U.S. patent application number 15/990491 was filed with the patent office on 2018-09-27 for device for determining a condition of an organ and method of operating the same.
The applicant listed for this patent is Nanyang Technological University, Singapore Health Services Pte Ltd. Invention is credited to Roger Beuerman, Shuo Chen, Quan Liu.
Application Number | 20180271368 15/990491 |
Document ID | / |
Family ID | 53524187 |
Filed Date | 2018-09-27 |
United States Patent
Application |
20180271368 |
Kind Code |
A1 |
Liu; Quan ; et al. |
September 27, 2018 |
DEVICE FOR DETERMINING A CONDITION OF AN ORGAN AND METHOD OF
OPERATING THE SAME
Abstract
In various embodiments, device for determining a condition of an
organ of either a human or an animal may be provided. The device
may include a first optical source and a second optical source. The
device may also include a detector. The device may additionally
include a lens system. The device may further include a switching
mechanism configured to switch between an optical examination mode
and a Raman mode. The lens system during the optical examination
mode may be configured to direct a first light emitted from the
first optical source. The lens system during the Raman mode may be
configured to direct a second light emitted from the second optical
source. The lens systems during the Raman mode may be further
configured to direct a third light to the detector.
Inventors: |
Liu; Quan; (Singapore,
SG) ; Beuerman; Roger; (Singapore, SG) ; Chen;
Shuo; (Singapore, SG) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
Nanyang Technological University
Singapore Health Services Pte Ltd |
Singapore
Singapore |
|
SG
SG |
|
|
Family ID: |
53524187 |
Appl. No.: |
15/990491 |
Filed: |
May 25, 2018 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
15111091 |
Jul 12, 2016 |
10004398 |
|
|
PCT/SG2014/000598 |
Dec 16, 2014 |
|
|
|
15990491 |
|
|
|
|
61926518 |
Jan 13, 2014 |
|
|
|
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
G02B 7/09 20130101; A61B
2576/02 20130101; A61B 5/7203 20130101; A61B 3/135 20130101; A61B
3/0075 20130101; A61B 3/14 20130101; A61B 5/0075 20130101; A61B
2503/40 20130101; A61B 3/0025 20130101 |
International
Class: |
A61B 3/135 20060101
A61B003/135; G02B 7/09 20060101 G02B007/09; A61B 3/00 20060101
A61B003/00; A61B 3/14 20060101 A61B003/14; A61B 5/00 20060101
A61B005/00 |
Claims
1. A device for determining a condition of an organ of either a
human or an animal, the device comprising: an optical source; a
detector; and a lens system; wherein the lens system is configured
to direct a light emitted from the optical source; and wherein the
lens system is further configured to direct a further light to the
detector.
2. The device according to claim 1, wherein the lens system
comprises an objective lens for focusing the light emitted from the
optical source.
3. The device according to claim 2, wherein the lens system further
comprises an actuator for controlling a position of the objective
lens.
4. The device according to claim 3, wherein the actuator is a
piezoelectric transducer.
5. The device according to claim 3, wherein the lens system
includes an actuator feedback circuit coupling the detector to the
actuator.
6. The device according to claim 5, wherein the actuator feedback
circuit is configured to receive an output from the detector and
further configured to provide a feedback to the actuator based on
the output from the detector.
7. The device according to claim 6, wherein the actuator feedback
circuit may be configured to determine a focus index based on the
output from the detector and further configured to provide a
feedback based on the determined focus index and a reference focus
index.
8. The device according to any of claims 1, further comprising: a
dynamic optical element for modulating the light emitted from the
optical source.
9. The device according to claim 8, wherein the lens system further
comprises a dynamic optical element feedback circuit coupling the
detector to the dynamic optical element.
10. The device according to claim 9, wherein the dynamic optical
element feedback circuit is configured to generate a skeletonized
line based on a line formed by the light.
11. The device according to claim 10, wherein the dynamic optical
element is configured to be adjusted based on a feedback from the
dynamic optical element feedback circuit until a focus index of
each pixel along a subsequent skeletonized line generated reaches a
maximum value.
12. The device according to claim 8, wherein the dynamic optical
element is a spatial light modulator or a digital micromirror
device.
13. The device according to claim 1, wherein the lens system
comprises a single beam splitter configured to direct the
light.
14. The device according to claim 1, further comprising: a
processor coupled to the detector.
15. The device according to claim 14, further comprising: one or
more filters configured to generate one or more narrow-band Raman
images from an image captured by the detector; wherein the
processor is configured to generate one or more reconstructed Raman
images based on the one or more narrow Raman images, each of the
one or more reconstructed Raman images corresponding to one
wavelength; and wherein the processor is further configured to
generate a Raman spectrum at each pixel based on the one or more
reconstructed Raman images.
16. The device according to claim 15, wherein the one or more
filters is configured to generate one or more reference narrow-band
Raman images from one or more reference images that contain full
spectral information at each pixel for all pixels; and wherein the
processor is configured to determine a Wiener matrix based on the
one or more reference narrow-band Raman images and the one or more
reference images.
17. The device according to claim 16, wherein the one or more
reference images is generated based on one or more reference
samples, each reference sample including one or more basic
biochemical components.
18. The device according to claim 16, wherein the processor is
configured to generate the one or more reconstructed Raman images
based on the one or more narrow-band Raman images and the Wiener
matrix.
19. The device according to claim 18, wherein the processor may be
configured to remove fluorescence background from the one or more
reconstructed Raman images.
20. The device according to 15, wherein the one or more narrow-band
Raman images may have a spectral resolution lower than the one or
more reconstructed Raman images.
21. The device according to 15, wherein the one or more filters may
be generated from one or more principal components based on Raman
spectra of the reference samples.
22. The device according to 1, further comprising: an interface
portion.
23. The device according to claim 22, wherein the lens system is
configured to direct the light emitted from the optical source to
the interface portion.
24. The device according to claim 23, wherein the lens system is
configured to direct the light from the interface portion to the
detector.
25. The device according to claim 1, wherein the further light has
a frequency shift from the light emitted from the optical
source.
26. The device according to claim 1, wherein the optical source is
a laser source.
Description
CROSS-REFERENCE TO RELATED APPLICATION
[0001] This application claims the benefit of priority of U.S.
patent application Ser. No. 61/926,518, filed 13 Jan. 2014, the
content of it being hereby incorporated by reference in its
entirety for all purposes.
TECHNICAL FIELD
[0002] Various aspects of this disclosure relate to devices for
determining a condition of an organ of either a human or an animal
as well as methods of operating the same.
BACKGROUND
[0003] Eye infection is a serious clinical problem, which has a
high likelihood to lead to blindness without proper treatment.
According to the Bulletin of the World Health Organization, eye
infection lead to about 1 million new cases of blindness in Asia
alone. One major cause of this serious disease is the infection
caused by microorganisms such as bacteria and fungi, which usually
occurs after intraocular surgery or simply is induced by a distant
infective source in the body. Early diagnosis is critical in the
management of this disease. The essential prerequisite for the
optimal treatment of eye infection is to identify the microorganism
causing infection as each type of microorganism causing eye
infection requires a different therapeutic approach. For example,
systemic antimicrobial therapy is usually recommended for patients
with endogenous endophthalmitis. In this case, the type and extent
of the infection needs to be diagnosed to determine potential
complications and find underlying systemic cause or risk
factors.
[0004] The current clinical procedure for identifying the
microorganism species causing eye infection includes Gram staining
and culture of aqueous and vitreous smear samples taken from the
surface of infected eyes, which is typically performed in the
pathology or microbiology department. Gram staining empirically
differentiates bacterial species into two large groups
(Gram-positive and Gram-negative) based on the chemical and
physical properties of their cell walls. It is fast and cheap.
However, it is not meant to be a definitive tool for diagnosis. For
example, it only works for bacteria and not every bacterium can be
definitively classified. Culture is considered as the gold standard
but this procedure is labor intensive and expensive. The cost for
Gram staining and culture can range from about 60 Singapore dollars
to about 180 Singapore dollars for material charge alone (excluding
labor), not mentioning a much larger cost incurred for disease
management if not treated in time and appropriately. It usually
takes a few days to culture the microorganisms in smears to get
reliable results. Such a long delay in diagnosis could result in
the exacerbation of patients' symptoms. The delay may also lead to
the optimal time frame for treatment being missed as well as the
subsequent rising cost for disease management. In addition, taking
smear samples from eyes for culturing is unpleasant and can be
challenging in some patients. In addition to Gram staining and
culturing, Polymerase Chain Reaction (PCR) is sometimes used to
assist diagnosis especially for those species that cannot be
cultured but PCR is in general expensive and its false-positive
rate is often high.
SUMMARY
[0005] In various embodiments, device for determining a condition
of an organ of either a human or an animal may be provided. The
device may include a first optical source and a second optical
source. The device may also include a detector. The device may
additionally include a lens system. The device may further include
a switching mechanism configured to switch between an optical
examination mode and a Raman mode. The lens system during the
optical examination mode may be configured to direct a first light
emitted from the first optical source. The lens system during the
Raman mode may be configured to direct a second light emitted from
the second optical source. The lens systems during the Raman mode
may be further configured to direct a third light to the various
embodiments, a method of operating a device for determining a
condition of an organ of either a human or an animal. The method
may include activating a switching mechanism to switch between an
optical examination mode and a Raman mode. During the optical
examination mode, a lens system may be configured to direct a first
light emitted from a first optical source. During the Raman mode,
the lens system may be configured to direct a second light emitted
from a second optical source. During the Raman mode, the lens
systems may be further configured to direct a third light to the
detector.
BRIEF DESCRIPTION OF THE DRAWINGS
[0006] The invention will be better understood with reference to
the detailed description when considered in conjunction with the
non-limiting examples and the accompanying drawings, in which:
[0007] FIG. 1 is a schematic illustrating a device for detecting a
condition of an organ of either a human or an animal according to
various embodiments.
[0008] FIG. 2 is a schematic illustrating a slit lamp system
according to various embodiments.
[0009] FIG. 3 is a schematic illustrating a device according to
various embodiments.
[0010] FIG. 4A is a schematic 400a of a setup to illustrate
focusing according to various embodiments.
[0011] FIG. 4B is an image of the setup illustrated in FIG. 4A
according to various embodiments.
[0012] FIG. 5A is a sequence of images obtained at the detector
according to various embodiments.
[0013] FIG. 5B is a plot of the focus index as a function of
distance in microns.
[0014] FIG. 6 is a schematic of a setup to illustrate auto-focusing
according to various embodiments.
[0015] FIG. 7 is a schematic of a setup to illustrate line
auto-focusing according to various embodiments.
[0016] FIG. 8A is a plot of intensity (arbitrary units) against
wavenumber (cm.sup.-1) illustrating the best case of the test
phantoms.
[0017] FIG. 8B is a plot of intensity (arbitrary units) against
wavenumber (cm.sup.-1) illustrating a typical case of the test
phantoms.
[0018] FIG. 8C is a plot of intensity (arbitrary units) against
wavenumber (cm.sup.-1) illustrating the worst case of the test
phantoms.
[0019] FIG. 9 is a schematic illustrating a procedure for Wiener
estimation according to various embodiments.
[0020] FIG. 10A is a plot of intensity (arbitrary units) against
wavenumber (cm.sup.-1) illustrating an original spontaneous Raman
data, including fluorescence background, of leukemia cells.
[0021] FIG. 10B is a plot of intensity (arbitrary units) against
wavenumber (cm.sup.-1) illustrating an original surface enhanced
Raman spectroscopy (SERS) data, including fluorescence background,
of blood serum sample.
[0022] FIG. 11A is a plot of intensity (arbitrary units) against
wavenumber (cm.sup.-1) illustrating measured spontaneous Raman
spectrum and the spontaneous Raman spectrum reconstructed by
traditional Wiener estimation for the best case using the best
combination of six commercial filters.
[0023] FIG. 11B is a plot of intensity (arbitrary units) against
wavenumber (cm.sup.-1) illustrating measured spontaneous Raman
spectrum and the spontaneous Raman spectrum reconstructed by
traditional Wiener estimation for a typical case using the best
combination of six commercial filters.
[0024] FIG. 11C is a plot of intensity (arbitrary units) against
wavenumber (cm.sup.-1) illustrating measured spontaneous Raman
spectrum and the spontaneous Raman spectrum reconstructed by
traditional Wiener estimation for the worst case using the best
combination of six commercial filters.
[0025] FIG. 11D is a plot of intensity (arbitrary units) against
wavenumber (cm.sup.-1) illustrating the transmittance spectra of
the six commercial filters corresponding to the typical case.
[0026] FIG. 12A is a plot of intensity (arbitrary units) against
wavenumber (cm.sup.-1) illustrating measured spontaneous Raman
spectrum and the spontaneous Raman spectrum reconstructed by
traditional Wiener estimation for the best case using the best
combination of six non-negative principal components (PCs) based
filters.
[0027] FIG. 12B is a plot of intensity (arbitrary units) against
wavenumber (cm.sup.-1) illustrating measured spontaneous Raman
spectrum and the spontaneous Raman spectrum reconstructed by
traditional Wiener estimation for a typical case using the best
combination of six non-negative principal components (PCs) based
filters.
[0028] FIG. 12C is a 1200c of intensity (arbitrary units) against
wavenumber (cm.sup.-1) illustrating measured spontaneous Raman
spectrum and the spontaneous Raman spectrum reconstructed by
traditional Wiener estimation for the worst case using the best
combination of six non-negative principal components (PCs) based
filters.
[0029] FIG. 12D is a plot of intensity (arbitrary units) against
wavenumber (cm.sup.-1) illustrating the transmittance spectra of
the six non-negative principal components (PCs) based filters
corresponding to the typical case.
[0030] FIG. 13A is a plot of intensity (arbitrary units) against
wavenumber (cm.sup.-1) illustrating measured surface enhanced Raman
spectroscopy (SERS) spectrum and the surface enhanced Raman
spectroscopy (SERS) Raman spectrum reconstructed by traditional
Wiener estimation for the best case using the best combination of
six commercial filters.
[0031] FIG. 13B is a plot of intensity (arbitrary units) against
wavenumber (cm.sup.-1) illustrating measured surface enhanced Raman
spectroscopy (SERS) spectrum and the surface enhanced Raman
spectroscopy (SERS) spectrum reconstructed by traditional Wiener
estimation for a typical case using the best combination of six
commercial filters.
[0032] FIG. 13C is a plot of intensity (arbitrary units) against
wavenumber (cm.sup.-1) illustrating measured surface enhanced Raman
spectroscopy (SERS) spectrum and the surface enhanced Raman
spectroscopy (SERS) spectrum reconstructed by traditional Wiener
estimation for the worst case using the best combination of six
commercial filters.
[0033] FIG. 13D is a plot of intensity (arbitrary units) against
wavenumber (cm.sup.-1) illustrating the transmittance spectra of
the six commercial filters corresponding to the typical case.
[0034] FIG. 14A is a plot of intensity (arbitrary units) against
wavenumber (cm.sup.-1) illustrating measured surface enhanced Raman
spectroscopy (SERS) spectrum and the surface enhanced Raman
spectroscopy (SERS) spectrum reconstructed by traditional Wiener
estimation for the best case using the best combination of six
non-negative principal components (PCs) based filters.
[0035] FIG. 14B is a plot of intensity (arbitrary units) against
wavenumber (cm.sup.-1) illustrating measured surface enhanced Raman
spectroscopy (SERS) spectrum and the surface enhanced Raman
spectroscopy (SERS) spectrum reconstructed by traditional Wiener
estimation for a typical case using the best combination of six
non-negative principal components (PCs) based filters.
[0036] FIG. 14C is a plot of intensity (arbitrary units) against
wavenumber (cm.sup.-1) illustrating measured surface enhanced Raman
spectroscopy (SERS) spectrum and the surface enhanced Raman
spectroscopy (SERS) spectrum reconstructed by traditional Wiener
estimation for the worst case using the best combination of six
non-negative principal components (PCs) based filters.
[0037] FIG. 14D is a plot of intensity (arbitrary units) against
wavenumber (cm.sup.-1) illustrating the transmittance spectra of
the six non-negative principal components (PCs) based filters
corresponding to the typical case.
[0038] FIG. 15 is a schematic illustrating a method of operating a
device for determining a condition of an organ of either a human or
an animal according to various embodiments.
[0039] FIG. 16 is a schematic of a device for determining a
condition of an organ of either a human or an animal according to
various embodiments.
DETAILED DESCRIPTION
[0040] The following detailed description refers to the
accompanying drawings that show, by way of illustration, specific
details and embodiments in which the invention may be
practiced.
[0041] The word "exemplary" is used herein to mean "serving as an
example, instance, or illustration". Any embodiment or design
described herein as "exemplary" is not necessarily to be construed
as preferred or advantageous over other embodiments or designs.
[0042] It should be understood that the terms "on", "over", "top",
"bottom", "down", "side", "back", "left", "right", "front",
"lateral", "side", "up", "down" etc., when used in the following
description are used for convenience and to aid understanding of
relative positions or directions, and not intended to limit the
orientation of any device, or structure or any part of any device
or structure.
[0043] The time consuming and costly culture procedure currently
used in clinical practice may warrant the development of a new
clinical method, which may be able to rapidly and accurately
identify the microorganism causing eye infection. Such a method may
assist an ophthalmologist to make appropriate therapeutic
strategies in a timely manner. Furthermore, it may be desirable to
eliminate the unpleasant step of getting smear samples from the eye
for the benefit of patients thus the method to be developed should
be able to scan the eye non-invasively. A device for use the such a
clinical method may be developed.
[0044] FIG. 1 is a schematic 100 illustrating a device for
detecting a condition of an organ of either a human or an animal
according to various embodiments. The device may include a first
optical source 102 and a second optical source 104. The device may
also include a detector 106. The device may additionally include a
lens system 108. The device may further include a switching
mechanism 110 configured to switch between an optical examination
mode and a Raman mode. The lens system 108 during the optical
examination mode may be configured to direct a first light emitted
from the first optical source 102. The lens system 108 during the
Raman mode may be configured to direct a second light emitted from
the second optical source 104. The lens systems 108 during the
Raman mode may be further configured to direct a third light to the
detector 106.
[0045] In other words, the device may be configured to operate in
two modes: an optical examination mode and a Raman mode. The device
may be configured to emit a first light from a first optical source
102 during the optical examination mode. The device may be
configured to emit a second light from a second optical source 104
and detect a third light using the detector 106 during the Raman
mode. The two modes may be switched between each other using a
switching mechanism 110.
[0046] The organ may be an eye. The device may be or may include an
opthalmoscope with a Raman module.
[0047] In various embodiments, the first optical source 102 may be
or may include an incoherent light source. The second optical
source 104 may be or may include a laser module.
[0048] The third light may be derived or may be based on the second
light. The second light emitted by second optical source 104 may
incident on the organ, e.g. an eye. The light reflected by the
organ may be the third light. Similarly, a first light emitted by
the first optical source 102 may incident on the organ to generate
or derive a fourth light. The fourth light may be the light
reflected by the organ when the first light is incident on the
organ.
[0049] The second light and the third light may be laser. The
second light emitted from the second optical source 104 may have a
frequency shift from the third light directed to the detector 106.
In other words, the second light may have a first frequency and the
third light may have a second frequency. The first frequency and
the second frequency may be different.
[0050] The first light and the fourth light may be incoherent
light. The electromagnetic waves making up the first light may not
have a constant phase difference and constant frequency. Similarly,
the electromagnetic waves making up the fourth light may not have a
constant phase difference and constant frequency. In various
alternate embodiments, the first light and the fourth light may
instead be coherent light or a laser and the first optical source
may be a coherent light source or a laser source, such as in a
laser opthalmoscope.
[0051] The device may further include an interface portion. The
interface portion may be a portion of the slit lamp system in which
the first light exits from the device and in which the fourth light
enters the device. During the optical examination mode, the lens
system 108 may be configured to direct the first light emitted from
the first optical source 102 to the interface portion. The first
light may be transmitted through the interface portion towards the
organ. The fourth light reflected from the organ may be transmitted
again through the interface portion. The fourth light may be in an
opposing direction to the first light. The device or the lens
system 108 may be configured to direct the fourth light to an
obsever such as an opthalmologist or a doctor. The observer may
thus be able to examine the organ. The device may include an
optical examination output portion such as an eye piece. The device
or the lens system 108 during the optical examination mode may be
further configured to direct the fourth light from the interface
portion to the optical examination output portion. The fourth light
may be derived from the first light. The fourth light may be
reflected by the organ when the first light is incident on the
organ. The fourth light may pass through the optical examination
output portion to the observer. The observer may examine the organ
by looking through or at the optical examination output
portion.
[0052] During the Raman examination mode, the lens system during
the Raman mode may be configured to direct the second light emitted
from the second optical source to the interface portion. The second
light may be transmitted through the interface portion towards the
organ. The third light reflected from the organ may be transmitted
again through the interface portion. The third light may be in an
opposing direction to the light. The device or the lens system 108
during the Raman mode may be configured to direct the third light
to the detector.
[0053] The lens system 108 may include an objective lens for
focusing the second light (onto the organ such as the eye). The
lens system 108 may further include an actuator for controlling a
position of the objective lens. By controlling the objective lens,
the actuator may control the focusing of the second light onto the
organ. The actuator is a piezoelectric transducer. The actuator may
move the objective lens upon application of a voltage to the
actuator. The lens system 108 may further include an actuator
feedback circuit coupling the detector to the actuator.
[0054] In various embodiments, a "circuit" may be understood as any
kind of a logic implementing entity, which may be special purpose
circuitry or a processor executing software stored in a memory,
firmware, or any combination thereof. Thus, in various embodiments,
a "circuit" may be a hard-wired logic circuit or a programmable
logic circuit such as a programmable processor, e.g. a
microprocessor (e.g. a Complex Instruction Set Computer (CISC)
processor or a Reduced Instruction Set Computer (RISC) processor).
A "circuit" may also be a processor executing software, e.g. any
kind of computer program, e.g. a computer program using a virtual
machine code such as e.g. Java.
[0055] The actuator feedback circuit may be configured to receive
an output from the detector and further configured to provide a
feedback to the actuator based on the output from the detector 106.
The actuator feedback circuit may allow for autofocusing. The
actuator feedback circuit may be configured to determine a focus
index based on the from the detector 106. The actuator feedback
circuit may be configured to determine a plurality of focus indexes
based on a plurality of outputs from the detector when the
objective lens is moved during calibration. The actuator feedback
circuit may be further configured to determine a maximum focus
index based on the plurality of focus indexes. The actuator may be
configured to move the objective lens during operation until the
index is at a reference focus index value, i.e. at the calibrated
maximum focus index. The focusing of the first light onto the organ
may be optimal when the focus index is at the predetermined
value.
[0056] In other words, the actuator feedback circuit may be
configured to determine a focus index based on the output from the
detector and may be further configured to provide a feedback based
on the determined focus index and a reference focus index. The
actuator feedback circuit may be further configured to control the
actuator to move the objective lens until the determined focus
index is substantially equal to the reference focus index.
[0057] The lens system may include a spatial light modulator (SLM)
for modulating the second light emitted from the second optical
source reflected (to the organ). In other words, the spatial light
modulator (SLM) is configured to reflect the second light from the
second optical source to the organ. In various alternate
embodiments, the lens system may include another dynamic optical
element, such as a digital micromirror device, for modulating the
second light emitted from the optical source reflected. The spatial
light modulator or dynamic optical element may be configured to
reflect different intensities of light. The lens system may further
include a spatial light modulator feedback circuit (or a dynamic
optical element feedback circuit) coupling the detector to the
spatial light modulator (or dynamic optical element). In various
embodiments, the spatial light modulator feedback circuit (or a
dynamic optical element feedback circuit) may be configured, e.g.
during operation, to generate a skeletonized line based on a line
formed by the second light (on the organ). The spatial light
modulator (or dynamic optical element) may be configured to be
adjusted based on a feedback, e.g. a feedback voltage, from the
spatial light modulator feedback circuit (or a dynamic optical
element feedback circuit) until a focus index of each pixel along a
subsequent skeletonized line generated reaches a maximum value. The
spatial light modulator (or dynamic optical element) may be
configured to reflect different intensities of light based on the
feedback. In other words, the lens system may include a dynamic
optical element for modulating the second light emitted from the
second optical source reflected (to the organ). The dynamic optical
element may be a spatial light modulator (SLM) or a digital
micromirror device. The lens system may further include a dynamic
optical element feedback circuit coupling the detector to the
dynamic optical element. The dynamic optical element feedback
circuit may be configured to generate a skeletonized line based on
a line formed by the second light
[0058] The maximum value may be predetermined, e.g. during a
calibration stage. During a calibration stage, the spacial light
modulator (or dynamic optical element) may be adjusted and a
plurality of focus indexes of each pixel may be determined. The
maximum value may be determined based on the plurality of focus
indexes of each pixel. During operation, the spatial light
modulator (or dynamic optical element) may be adjusted until the
maximum value is reached.
[0059] The lens system 108 may include one or more beam splitters
or dichroic mirrors configured to direct the first light during the
optical examination mode and further configured to direct the
second light during the Raman mode. In other words, the one or more
beam splitters or dichroic mirrors may be used in both modes, i.e.
shared for both modes.
[0060] The device may further include a processor coupled to the
detector 106. The or lens system 108 may additionally include one
or more filters configured to generate one or more narrow-band
Raman images from an image (of the organ) captured by the detector
106. The processor may be configured to generate one or more
reconstructed Raman images based on the one or more narrow Raman
images. Each of the one or more reconstructed images may correspond
to one wavelength or one range of wavelengths. The processor may be
further configured to generate a Raman spectrum at each pixel based
on the one or more reconstructed Raman images. There may be a
plurality, e.g. hundreds, of reconstructed images, each
corresponding to one wavelength (or wavenumber). The intensity
values of these reconstructed images may be concatenated to form a
Raman spectrum at each pixel.
[0061] The one or more filters may be configured to generate one or
more reference narrow-band Raman images from one or more reference
images, for instance, during the calibration stage. The one or more
reference images may contain or include full spectral information
at each pixel (for all pixels). The processor may be further
configured to determine a Wiener matrix based on the one or more
reference narrow-band Raman images and the one or more reference
images. The one or more reference images may be generated based on
the one or more reference samples. Each reference sample may
include one or more basic (biochemical) components. The processor
may be configured to generate the one or more reconstructed Raman
images based on the one or more narrow-band Raman images and the
Wiener matrix. The processor may be configured to remove
fluorescence background from the one or more reconstructed Raman
images. The one or more narrow-band Raman images may have a
spectral resolution lower than the one or more reconstructed Raman
images. The one or more filters may include one or more principal
component filters. Additionally or alternatively, the one or more
filters may include one or more commercial filters and/or one or
more gaussian filters. The one or more filters may be generated
from one or more principal components based on or calculated from
Raman spectra of the reference portion.
[0062] FIG. 2 is a schematic 200 illustrating a slit lamp system
according to various embodiments. The slit lamp may be configured
to determine a condition of an organ such as an eye 214. The slit
lamp system may include an optical source 202 and a lens system
208. The lens system may be configured to direct a first light
(indicated by 218a) emitted from the first optical source 202 to
the eye 214. A fourth light (indicated by 218b) may be reflected
from the eye 214.
[0063] The slit lamp system may further include an interface
portion 210. The interface portion 210 may provide an interface
between the slit lamp system and the eye 214. The interface portion
210 may be a portion of the slit lamp system in which the first
light exits from the slit lamp system and in which the fourth light
enters the slit lamp system. The lens system 208 may be configured
to direct the first light emitted from the first optical source 202
to the interface portion 210. The first light may be transmitted
through the interface portion 210 towards the eye 214. The fourth
light reflected from the organ may be transmitted again through the
interface portion 210. The fourth light may be in an opposing
direction to the first light. The lens system 208 may be configured
to direct the fourth light to an obsever such as an opthalmologist
or a doctor. The observer may thus be able to examine the organ.
The slit lamp system may include an optical examination output
portion 212 such as an eye piece. The lens system 208 may be
further configured to direct the fourth light, the fourth light
derived from the first light, from the interface portion 210 to the
optical examination output portion 212. The fourth light may pass
through the optical examination output portion 212 to the observer.
The observer may examine the organ by looking through or at the
optical examination output portion. The lens system 208 may include
one or more beam splitters or dichoric mirrors 216. The lens system
208 may further include other optical components for directing the
first light and/or the second light.
[0064] FIG. 3 is a schematic 300 illustrating a device according to
various embodiments. The device may be modified from the slit lamp
system illustrated in FIG. 2. The device may include the slit lamp
system illustrated in FIG. 2. The device may further include a
Raman module.
[0065] Various embodiments may provide a device for eye scanning
based on Raman spectroscopy which aims to rapidly and noninvasively
detect Raman spectra from infected cornea and identify the species
of microorganisms causing eye infection. The device may include a
Raman module in a slit lamp ophthalmoscope so that the scanning
procedure and the outlook of the equipment are similar to those in
a routine slit lamp examination. The observer or operator may be
able to conveniently switch between an optical examination mode
(also referred to as slit lamp examination mode) and a Raman mode
(using the switching mechanism) while the imaged area (of the eye)
remains unchanged. Detected Raman spectra may be processed by a
method of multi-variate statistical analysis to identify the
species of microorganisms causing eye infection. The outcome of
Raman analysis may assist the clinician in diagnosing eye
infection. The device may serve as an adjunct tool to provide an
alternative to the current clinical procedure for the diagnosis of
eye infection in the short term. Based on the result of rapid Raman
analysis, the clinician may decide whether the culture step is
necessary to reconfirm the diagnosis and/or make appropriate
treatment plans early. The wide use of this technique may reduce
the need of the expensive and time consuming culture step in the
current procedure and cut down the cost of eye infection
management.
[0066] The device may include a first optical source 302 and a
second optical source 304. The device may also include a detector
306. The device may additionally include a lens system 308. The
device may further include a switching mechanism configured to
switch between an optical examination mode and a Raman mode. The
lens system 308 during optical examination mode may be configured
to direct a first light emitted from the first optical source 302.
The lens system 308 during the Raman mode may be configured to
direct a second light emitted from the second optical source 304.
The lens systems 308 during the Raman mode may be further
configured to direct a third light to the detector 306.
[0067] The operation of the device during the optical examination
mode may be similar to the operation of the slit lamp system
illustrated in FIG. 2. The device may include an interface portion
310 and an optical examination output portion 312.
[0068] The interface portion 310 may be a portion of the device in
which the first light exits from the device and in which the fourth
light enters the device. The lens system 308 may be configured to
direct the first light emitted from the first optical source 302 to
the interface portion 310. The first light may be transmitted
through the interface portion 310 towards the eye 314. The fourth
light reflected from the eye 314 may be transmitted again through
the interface portion 310. The lens system 308 may be configured to
direct the fourth light to an obsever such as an opthalmologist or
a doctor through the optical examination output portion 312. The
lens system 308 may be further configured to direct the fourth
light, the fourth light derived from the first light, from the
interface portion 310 to the optical examination output portion
312.
[0069] During the Raman examination mode, the lens system 308
during the Raman mode may be configured to direct the second light
(indicated by 318a) emitted from the second optical source to the
interface portion 310. The second light may be transmitted through
the interface portion towards the eye 314. The third light
(indicated by 318b) reflected from the eye 314 may be transmitted
again through the interface portion 310. third light may be in an
opposing direction to the second light. The device or the lens
system 308 during the Raman mode may be configured to direct the
third light to the detector 306. The interface portion 310 may be a
portion of the device in which the second light exits from the
device and in which the third light enters the device.
[0070] The device may further include a function generator 320a and
a delay generator 320b. The arrows 322a, b, c may represent the
flow of control signals. The function generator 320a may be
configured to provide control signal 322a to the second optical
source 304 to activate the second optical source 304. The function
generator 320b may be configured to provide control signal 322b to
the delay generator 320b. The delay generator 320b may be
configured to provide control signal 322c to the detector 306 after
a predetermined delay from receiving control signal 322b.
[0071] The lens system 308 may include one or more beam splitters
or dichoric mirrors 316a-c. The one or more beam splitters or
dichroic mirrors 316a-c may be configured to direct the first light
during the optical examination mode and further configured to
direct the second light during the Raman mode. In other words, the
one or more beam splitters or dichroic mirrors may be used in both
modes, i.e. shared for both modes. The dichroic mirrors 316a-c may
be separate mirrors.
[0072] The lens system 308 may further include other optical
components for directing the first light and/or the second
light.
[0073] Various embodiments may provide fast and accurate Raman
measurements as required by clinical examination. The integration
of a Raman module into a slit lamp ophthalmoscope may require
careful detailed design to achieve fast and accurate Raman
measurements as required by clinical examination which is highly
challenging.
[0074] For instance, it may not practical to expect a clinician to
be able to manually focus laser light onto the area of interest on
the cornea surface considering that they are not experts in optical
alignment. The fact that the cornea is transparent may make it more
difficult to achieve a good focus. Unfortunately, this may be
required to get a good Raman signal.
[0075] Various embodiments may include an autofocusing method and
system developed to eliminate or reduce the need of manual
focusing. A red laser adjusted at a small power following exactly
the same optical path as the excitation light for Raman excitation
may be used to facilitate alignment. The laser spot for alignment,
which may be visible to clinicians, may be manually moved
vertically along the line illuminated by the slit lamp to help the
physician trace the location to be examined by Raman measurements.
Once the location of interest is identified, the autofocusing
procedure may be started. The procedure may take advantage of the
fact that the reflected light intensity increases dramatically when
a laser spot crosses a boundary with refractive index mismatch. A
computer may control a piezoelectric transducer (PZT) based
actuator to move a microscope objective lens to achieve
autofocusing.
[0076] The same light source for Raman excitation (working at a
small power) and detector 306 (e.g. spectrograph and charged
coupled device (CCD)) may also be used for alignment to save the
extra light source and photodetector for alignment. The
disadvantage of this approach may be that the clinician may not be
able to see the laser spot as clear as the red laser spot. Only the
vertical dimension of the image in this case may represent the
spatial dimension that is used to calculate the contrast and spot
size. The other dimension (horizontal dimension) may correspond to
the spectral dimension and should not be used for the purpose of
calculating contrast and spot size.
[0077] Another issue may be that a patient's eye may not stay at
one fixed position for long, typically only a couple of seconds or
shorter. It may be challenging to detect Raman spectra with a
decent signal to noise ratio within such a short time frame.
[0078] The eye movement problem may be overcome by fast data
acquisition, ideally in real time. Within the short data
acquisition period, the cornea may be viewed as stationary and the
effect of eye movement on Raman spectra may be neglected. To
achieve this, a modified Wiener estimation for spectral
reconstruction and relevant algorithms may be used to speed up data
acquisition. The method of data acquisition using the modified
Wiener estimation may be distinguished from most other Raman
systems. A spectrograph with much poorer spectral resolution
compared to a normal one used in Raman acquisition may be used to
improve the signal to noise ratio by providing a larger bandwidth
at each wavenumber. Then a modified Wiener estimation may be used
to reconstruct Raman spectra at the required spectral resolution
rapidly.
[0079] The idea of using Raman spectroscopy to identify the
microorganisms causing eye infection may be based on the following
phenomena. First, bacteria and fungi, which are the two major
microorganisms causing infection, may exhibit unique Raman
fingerprints. Second, the change in the biochemical composition,
and thus the Raman patterns, of ocular tissues induced by eye
infection may vary with the microorganism species. This may be a
secondary effect compared to the Raman spectra of the microorganism
in terms of the diagnostic value. Raman spectroscopy has been
recently used for the differential diagnosis of eye infection and
uveitis in rabbit iris in vitro and monitoring intraocular drugs
against endophthalmitis, which demonstrates the feasibility of
using Raman spectroscopy for the in vivo identification of
microorganisms causing eye infection.
[0080] However, the following requirements may have to be fulfilled
in order to perform Raman measurements in the eye in vivo.
[0081] 1. The excitation power density may be required to be lower
than the safety threshold prescribed in International Laser Safety
Standards while data acquisition may need to be fast to prevent the
motion artifact.
[0082] 2. There may be a requirement to avoid optical alignment or
realignment when switching between the route slit lamp examination
mode and Raman mode.
[0083] 3. Data analysis may need to be fast to provide quick
feedback.
[0084] The strategies below may be employed to address these
requirements one by one.
[0085] 1. A strategy may be taken to improve the signal to noise
ratio in order to achieve the goals of lowering the excitation
power density and fast data acquisition. Various filters may be
used to remove the sideband in the excitation light and the
influence of ambient light on Raman spectra. A lock-in detection
scheme using a gain-modulated intensified CCD may be utilized to
improve the signal-to-noise ratio.
[0086] 2. The Raman module may be designed to minimize the changes
or modification to the slit lamp system as shown in FIG. 2A. FIG.
2B is a non-limiting example to illustrate one of several options
for incorporating the Raman module. A computer code may be
developed to automate the operation of switching between the
routine slit lamp examination mode and Raman mode. In other words,
the switching mechanism may include a processing circuit including
a computer algorithm. A clinician may work in the routine
examination mode first to locate the target area. Then the scanner
may be switched to Raman mode to take Raman spectra from the area
without the need of any further adjustment.
[0087] 3. An ex vivo study may be carried out to identify optimal
Raman bands for differentiating various species of microorganisms.
Only selected Raman bands may be involved in clinical data analysis
to speed up the diagnosis.
[0088] By applying these strategies, a device including a Raman
module integrated with a slit lamp system may enable fast and
sensitive Raman measurements from the eye 314 without interrupting
the routine slit lamp examination procedure. The high sensitivity
of Raman measurements to a range of microorganisms causing
infection demonstrated previously may help ensure the accuracy of
the non-invasive optical diagnosis. With the advance in laser
technology and sensitive optical detectors, the cost of optical
components in a sensitive Raman system has dropped significantly,
which may make it feasible to build a cost-effective Raman module
for eye scanning. Due to the nature of non-contacting optical
measurements, the device may require minimum maintenance, thus
further bringing down the total cost of operating such a system on
a regular basis.
[0089] The lens system 308 may include an objective lens (not shown
in FIG. 3) for focusing the second light (onto the organ such as
the eye 314). The lens system 308 may further include an actuator
(not shown in FIG. 3) for controlling a position of the objective
lens. By controlling the objective lens, the actuator may control
the focusing of the second light onto the eye 314. The actuator is
a piezoelectric transducer. The actuator may move the objective
lens upon application of a voltage to the actuator. The lens system
308 may further include an actuator feedback circuit (not shown in
FIG. 3) coupling the detector to the actuator.
[0090] The actuator feedback circuit may be configured to receive
an output from the detector 306 and further configured to provide a
feedback, e.g. a feedback voltage, to the actuator based on the
output from the detector 306. The actuator feedback circuit may
allow for autofocusing. The actuator feedback circuit may be
configured to determine a focus index based on the output from the
detector 306. The actuator feedback circuit may be configured to
determine a plurality of focus indexes based on a plurality of
outputs the detector when the objective lens is moved during
calibration. The actuator feedback circuit may be further
configured to determine a maximum focus index based on the
plurality of focus indexes. The actuator may be configured to move
the objective lens during operation until the focus index is at a
reference focus index value, i.e. at the calibrated maximum focus
index. The focusing of the first light onto the eye 314 may be
optimal when the focus index is at the predetermined value.
[0091] In other words, the actuator feedback circuit may be
configured to determine a focus index based on the output from the
detector 306 and may be further configured to provide a feedback,
e.g. a feedback voltage, based on the determined focus index and a
reference focus index. The actuator feedback circuit may be further
configured to control the actuator to move the objective lens until
the determined focus index is substantially equal to the reference
focus index.
[0092] FIG. 4A is a schematic 400a of a setup to illustrate
focusing according to various embodiments. The device according to
various embodiments may include various components of the setup in
a similar manner. In various embodiments, references to the setup
may include references to the device. FIG. 4B is an image 400b of
the setup illustrated in FIG. 4A according to various
embodiments.
[0093] The auto-focusing option may include two portions 1)
hardware components and/or 2) software. The orientation of the
experimental setup for focusing the specific area of the sample
includes both hardware and software components. The hardware
components used for the experimental setup may be illustrated in
FIG. 4A. The hardware component may be used to obtain the stack of
images and the software may be used to perform the image
processing.
[0094] The setup may include a laser (or LASER) source 404. The
laser source may also be used in the device. LASER is an acronym
for light amplification by stimulated of radiation. A laser is a
component which may be configured to emit light by the of
stimulated emission. When a particle is hit by the photons it may
absorb some energy and may jump to the excited state from the
ground state. When returning back to the original position, the
particle may emit some of its energy in the form of photons. Lasers
may be used in many biomedical and biological applications due to
different
[0095] The setup may also include a beam splitter 416. The beam
splitter 416 may also be included in the device. As the name
implies, the beam splitter 416 may split an optical beam into two
by allowing a first light to pass through it and a second light to
be reflected at substantially 90.degree. (at the point of
incidence). The optical beam may be treated as the second light as
highlighted earlier. The beam splitter 416 may work in a similar
manner to a mirror in transmitting the part of incident light (the
second light). One part of the second light may be made to pass
through the beam splitter 416 and the rest of the second light may
be reflected from the reflecting surface of the beam splitter
416.
[0096] The setup and/or device may include one or more objective
lens 424a, 424b. For the optical imaging, microscopic objective may
play a major role in the determining the image quality and may also
interpret the primary image formation process. The primary purpose
of objective lens 424b may be to collect the light from the sample
or object 414 and to magnify the information and to provide the
magnified information to detector 406. Objective lens 424b may come
with different degrees of magnification power. The sample or object
414 may be an organ such as an eye but may also be a non-living
object for the purposes of experiment. The third light reflected
from the object 414 may pass through the beam splitter 416 to the
detector 406.
[0097] Objective lens 424a, 424b may be characterized by two
parameters, namely magnification factor and numerical aperture. The
objective lens 424a, 424b may provide information by enlarging the
content with a specified range. The magnification factor for
objective lens may range from 40.times. to 100.times.. The
numerical aperture may be defined as the accepted angle of the lens
424a, 424b from which it is determined how the lens 424a, may
readily emit or accept light. As numerical aperture increases, the
working distance may decrease. The working distance may be defined
as the maximum distance between sample 414 and front part of the
lens 424b from which all the information of the sample be collected
and it is defined as sharp focus. The parfocal length may be
defined as the distance between the objective mounting position to
the sample surface, i.e. the of focal length and working
distance.
[0098] The setup or device may also include a detector such as a
charge coupled device (CCD) 406. The detector may include a
photographic film. The CCD may include a thin silicon wafer which
may be divided into a plurality of small light sensitive areas.
Each separate area or square may be referred to as a photosite.
Here each photosite may be substantially equivalent to a pixel of
the image. Each photosite may include a capacitor which is
positively charged. Basically, the CCD is an analog device which
may convert light to electrons by photoelectric effect. Since the
photosites are positively charged, the electrons, which are
negatively charged, may be attracted towards photosites. The
particular movement of charges inside the device may provide the
output voltage, which may be proportional to the number of photons
that are incident on the photosites. However, the analog signal may
be converted into digital signals. The CCD may record the video
instead of taking the pictures. The cost of the scientific CCD may
be expensive because the size is big and only little silicon wafers
may fit to design.
[0099] Exposure may start at the time the capacitors of the
photosites are positively charged and may end by disconnection and
opening the shutter of the CCD. The light from the objective lens
424b may be made to pass through the silicon within the CCD.
However this may lead to transfer of some electrons from low energy
valence band to energy conduction band. Some of the electrons may
be attracted towards the positively charged capacitor, which may
allow the capacitor to discharge partially. The amount of discharge
may be directly equal or may be proportional to the number of
photons incident on the photosite during the exposure time. At or
near the end of the exposure, the at each photosite may be
amplified and may pass through the analog to digital converter
(ADC) device to digitize the signal.
[0100] While a piezoelectric transducer (PZT) has not been used in
the preliminary experiment, a device including a piezoelectric
transducer may be envisioned. A piezoelectric transducer may be a
device which converts mechanical movement to electrical energy and
electrical energy into mechanical movement. The PZT may include a
polarized material. When the electric current passes through the
transducer, the polarized material may realign in the different
manner compared to an initial alignment, consequently producing a
different shape of the material and generating mechanical movement.
This process may be called electrostriction.
[0101] A fully polarized material such as quartz may generate
electric energy according to the change in dimension of the
material. This process may be called piezoelectric effect.
[0102] The polarized material may be or may include a ceramic
material, which may have a high efficiency to change size and
shape. The efficiency of the transducer may be measured by the
ratio of output energy to input energy. The efficiency of the
transducer may be good when the output energy is greater than the
input energy.
[0103] The principle of focusing technique may depend on the image
spatial resolution. For the digital images, resolution may be
dependent on the number of pixels in an image. The spatial
resolution is used to find closeness of pixels revolved to create
an image. Even if the number of pixels is high, the spatial
resolution of the image may not be good. The spatial resolution
depends on the clarity of the image.
[0104] When the image is in focus, the spatial resolution may be
high. Conversely, when the image is out of focus, the spatial
resolution may be lower. The determination of the focus position
may be widely dependent on the spatial frequency at different
planes of each image extracted.
[0105] It may be important to align the laser beam straight because
even with the slight disturbance, it may lead to misalignment of
the images at different distance. The components shown in FIG. 4
are aligned in such way that the laser beam travels all through the
components in the straight direction without any deviation.
[0106] The laser beam from the laser diode 404 may be first made to
pass through the infinite microscopic objective lens 424a. Without
using the microscopic objective 424a in the preliminary setup of
the experiment, it was found that the images are not along the same
point, i.e. the image obtained was fluctuating along the frame with
respect to different distances. The microscopic objective lens 424a
(M1) was included to minimize the fluctuation and to make the size
of the beam larger. The pinhole 426a and the plano convex lens 426b
may be placed in between the objective 424a and the mirror 428 to
collimate the forecoming light in order to prevent it from
diverging.
[0107] Later the beam is reflected from the mirror 428 for the sake
of requirement of the experiment. As already mentioned, a beam
splitter 416 may be included to split the light into two. The beam
splitter 416 may reflect about 50% of light to the microscopic
objective 424b (M2) and allow another about 50% of light to pass
through in a straight direction (not shown in FIG. 4A). The
microscopic objective 424b may play a major role in determining the
quality of the image and to trace the focus position. In initial
experiment, objective lens 424b is placed on the manual translator
430, which moves the objective lens 424b forward and backward along
a direction, e.g. in the z direction, relative to the sample 414.
The objective lens 424b may be moved relative to the sample 414 to
achieve the best focusing.
[0108] The manual translator 430 may have a millimeter range along
the main scale and a microns range along the rotating scale. The
light from the objective lens 424b M2 may converge at some point
when moving out of the lens 424b. The converging distance from the
particular point to the front of the lens 424b may be called the
working distance of the objective lens 424b. The point, i.e. the
focus point, may be related to the working distance of the
objective lens 424b. The objective lens used here may be purchased
from Thorlabs. Here the microscopic objective used has the
magnification factor of 40.times. with a working distance is about
0.6 mm. The effective focal length of the objective lens 424b is
4.5 mm and parfocal length is about 45.06 mm. The numerical
aperture of the objective lens 424b is 0.65.
[0109] The reflected light (i.e. third light) from the sample 414
may pass through the beam splitter 416. The reflected light may
contain the collected information of the sample and the detector
406 may capture the image. Tube lens 432a and microscopic eye piece
lens 432b may be placed between the beam splitter 416 and CCD 406.
The purpose of the objective lens 424b is to collect the image of
the sample 414 at infinity. The objective lens 424b sends reflected
light from the sample 414 as a bundle of parallel lights across to
tube lens 432a. The tube lens 432a behaves as a receiver and sender
which may center the parallel lights from the lens 424b to the
centre part of the detector 406. Microscopic eye piece lens 432b
may collect the light from tube lens 432a so that the lights from
lens 432b are again substantially parallel. Tube lens 432a may be
accompanied with the microscopic eye piece lens 432b for obtaining
better results. The main advantage of using the tube lens 432a is
that it provides a space between beam splitter 416 and detector 406
so an external optical component like another beam splitter or
filter may be included when necessary.
[0110] The working distance of the tube lens 432a may vary
according to different manufacturers. The tube lens may have a
working distance range from about 70 to about 200 mm. The focal
distance of the tube lens 432a may be about 200 mm. The tube lens
432a may be placed at a distance from about 70 mm to about 200 mm
from the objective lens 424b. When the tube lens 432a is placed at
below 70 mm from the objective lens the resultant image may be
affected by aberrations. Conversely, when the tube lens 432a placed
at beyond 200 mm from the objective lens 424b, the scan lens in the
tube may be overfilled, which may result in inaccurate results.
[0111] The macroscopic eye piece lens 432b may have a different
magnification power from tube lens 432a. The main purpose of the
macroscopic eye piece lens 432b is magnification. The magnification
factor of 10.times. is used in our experiment. The light reflected
from the sample 414 is magnified accordingly and fed to the central
part of the detector 406.
[0112] The detector 406 may be an Electron Multiplying Charge
Coupled Device (EMCCD). The manufacture of the EMCCD used in our
study is the Princeton Instruments. ProEM cameras are designed in
such a way to overcome the challenges of low-light, high frame
rate, and light-starved applications. A ProEM camera may include
512.times.512 back-illuminated EMCCD and may support both electron
multiplication (EM) and traditional readout ports. The images of
the sample 414 may be obtained by varying the distance of the
sample 414 with the detector 406 at regular intervals.
[0113] A focusing technique based on the image spatial resolution
may be provided. As discussed earlier, the spatial resolution is a
factor which determines how closely the are related to form an
image. When the objective lens 424b M2 is in focus position, the
image obtained at that position may have a high value of the
spatial resolution. When objective lens 424b M2 moves out of the
focus position, the spatial resolution of the obtained may
decrease. Consequently, when the image spatial resolution
decreases, the high frequency components of the image may also be
decreased. The device or setup may be configured to measure the
high frequency components for the captured images at each plane and
determine an optimal focus.
[0114] Generally, the high frequency content is extracted from the
full frequency spectrum of the captured image by using a high band
pass filter. The external analog filters or the digital filters
inside the computer may be used for this purpose. In other words,
the filter may be a physical filter or a digital filter.
[0115] For each turn, the manual translator 430 may be moved at
constant intervals of distance. The stack of images with respect to
the change in distance may be captured by the detector 406. With
the help of the computer, digital image processing may be carried
out for all the images to find the focus index. The focus index may
be described as the ratio of the sum of square of the each pixel
value of the convoluted image to the square of the sum of the pixel
values of the original image. The convoluted image is may be
generated by convolution of the image with the high pass
filter.
[0116] The focus index may be calculated from the formula stated
below,
F ( z ) = x y [ f ( x , y ) i 2 ( x , y ) ] 2 [ x y i z ( x , y ) ]
2 ( 1 ) ##EQU00001##
where x, y are the index of the pixels, i.sub.z(x,y) denotes the
value of each pixels, f(x,y) is the value of high pass filter and
{circle around (x)} operator denotes convolution factor
[0117] The added advantage of using the digital filters is that
using the digital filters may provide a wide number of choices to
select the different filters. In the experiment, the kernel integer
filters are used for the image processing. The kernel filters are
the most used filters for convoluting the image. Kernel filters may
provide several options like blurring, edging, sharp detection,
smoothening and even more in the field of image processing.
Generally, kernel filters may allow both low pass and high pass
filtering.
[0118] The sample or object 414 may be a glass slide with particles
distributed on the glass slide. Each particle is about 10 .mu.m in
size. The sample particle on the glass slide may be placed in front
of the objective lens 424b (M2). The manual translator 430 may be
moved along a line (i.e. one dimensional translation) towards and
away from the sample or object 414. The objective lens 424b may be
placed as near as possible the sample or object 414 initially but
without touching the glass slide. The objective lens is sensitive
component and even a small disturbance can cause damage to lens,
one should be very careful with moving the objective lens 424b near
the sample or object 414. The objective lens 424b may be moved away
from the sample or object 414 using the .
[0119] The focus index may be calculated for each image on the
detector 406 according the formula. The first image may taken when
the objective is placed very near to the sample and the later
images may be obtained when moving the objective lens away from the
sample which is mounted on the manual translator 430.
[0120] FIG. 5A is a sequence 500a of images obtained at the
detector 406 according to various embodiments. The images are
obtained by varying distances of the objective lens 424b from
sample or object 414 (captured with constant intervals). The images
are obtained in the interval of 0.5 microns distance. The resultant
images were originally in the tif format and it was converted to
bmp file with the help of Image J software. The images obtained was
read by the computer and stored in the appropriate space for
subsequent use in the program. MATLAB software was used for the
image processing and to perform calculations. Any other software
like C, C++ may also be used. The advantages of MATLAB may include
that the speed is high and hence may result in a less time
consuming process. Manual adjustment may be time consuming and
automation may increase the speed.
[0121] The integer kernel filter may be used as the high pass
filter to filter the high frequency content of the images from the
full spectrum of the images. The integer filter
used here is,
1 - 2 - 2 - 1 - 12 - 1 - 1 - 2 - 1 ( 2 ) ##EQU00002##
[0122] As discussed, the first image is taken when the objective is
placed very near to the sample. The region of interest (ROI), i.e.
where the particle is spread vastly on the glass slide, is found on
the first image. The objective lens 424b is securely fixed to the
translator 430 for reducing errors. The manual translator 430 is
moved away from the sample 414. The objective lens 422b is moved
away from the sample 414 by moving of the manual translator.
[0123] Further images are captured at subsequent constant
intervals. The total illumination of the original image is
calculated by summing up of all pixel values in each images and the
whole sum is squared according to Equation (1). The convoluted
image is generated by convoluting the image with the filter in
(2).
[0124] Then the sum of square of all the gray pixels values (for
the convoluted image) is calculated. The focus index for each image
is then calculated by dividing the numerator, i.e. the square of
the sum of the original image, with the denominator, i.e. the sum
of the square of the pixel value of the convoluted image. The focus
index values may be plotted as a function of the distance of the
objective lens 424b from the object 414. FIG. 5B is a plot 500b of
the focus index as a function of distance in microns.
[0125] A bell shaped curve may be obtained as shown in FIG. 5B. The
peak in the curve provides the highest value of the focus index.
This value may represent a focus image at detector 406 when the
objective lens 424b is at an optimal distance from the object 414.
The high frequency content of the focus image may be high compared
to the other images which are out of focus.
[0126] The eye is a semi-transparent object which has many layers
with different values of refractive index. As such, the experiment
has been done on a transparent surface. In summary, a stack of
images may be generated at the detector 406 and the focus index of
each image may be determined. The highest value of the focus index
may be determined from the focus indexes of the images. The image
with the highest focus index value may be the focus image. The
spatial frequency of each image may correspond to the spatial
resolution of the image.
[0127] FIG. 6 is a schematic 600 of a setup to illustrate
auto-focusing according to various embodiments. The setup may be
similar to the setup illustrated in FIG. 4A but with the manual
translator replaced by a piezoelectric transducer 630 (such as a
lead zirconium titanate (PZT) transducer) and an actuator feedback
circuit 634 coupled between the detector 406 and the piezoelectric
transducer 630. The device according to various embodiments may
include various components of the setup in a similar manner. In
various embodiments, references to the setup may include references
to the device.
[0128] The actuator feedback circuit 634 may be configured to
receive an output from the detector 406 and further configured to
provide a feedback, e.g. a feedback voltage, to to the
piezoelectric transducer 630 based on the output from the detector
406. The first calculated focus index for an image may be taken as
the reference value (reference focus index). The feedback voltage
may be provided to the piezoelectric transducer 630 for an
subsequent image based on the focus index of the subsequent image.
The feedback voltage may be further based on the reference focus
index or the focus index of a preceding image. The piezoelectric
transducer 630 may be moved accordingly towards or away from the
object 414, e.g. the eye, until the focus index reaches a maximum
value, i.e. when the focus image is detected. The actuator feedback
circuit 634 may be configured to determine a focus index based on
the output from the detector 406 and may be further configured to
provide a feedback, e.g. a feedback voltage, based on the
determined focus index and a reference focus index.
[0129] FIG. 7 is a schematic 700 of a setup to illustrate line
auto-focusing according to various embodiments. In various
embodiments, references to the setup may include references to the
device.
[0130] The setup may include a laser source 704 and a detector 706
such as a spectrometer. The spectrometer may include a CCD. The
setup may be configured to focus a light (i.e. the second light
represented by 718a) on a sample 714. The sample 714 may be an
organ such as an eye but may be a non-living object for the
purposes of the experiment. The setup may be further configured to
direct a third light from the sample 714 to the detector 706. The
third light may be the second light reflected from the sample
[0131] The setup may further include a spatial light modulator
(SLM) 736 for modulating the second light emitted from the laser
source 704. The setup may further include a cylindrical lens (CL)
738 so that the second light focused onto the sample is a curved
line.
[0132] Most current autofocusing methods may be designed for
focusing on a single point. However, for Raman measurements on the
eye surface, line scanning may be advantageous over point scanning
because the former method offers much higher speed in data
acquisition. There may be some difficulties involved in line
autofocusing for the purposes of eye scanning. One difficulty may
be that there are currently no existing methods for autofocusing on
a line. Another difficulty may be that the eye surface is curved.
The second difficulty may mean that light needs to be focused on a
curve instead of a straight line. The setup may include a
cylindrical lens (CL) 738 and a spatial light modulator 736 to
implement line autofocusing as shown in FIG. 7. The setup may also
include a filter 740, e.g. a long-pass filter (LP), between the
detector 706 and a dichroic filter (DF) 716.
[0133] Light (i.e. the second light) from the laser source 704 may
be expanded by the beam expander (BE) 740 first. The light (i.e.
the second light) may be deflected by the SLM 736 and mirror 728
(and through the dichroic filter (DF) 716) onto the cylindrical
lens (CL) 738. After passing through the cylindrical lens (CL) 738
and the objective lens (OBJ), the light (i.e. the second light) may
form a line focus onto the surface of a sample to yield a bright
line. The line may be a curve if the sample surface is curved. The
line may be distorted if part of the light is not focused well. The
line formed on the tissue surface may be imaged by the spectrometer
706 when the spectrograph inside the spectrometer is set to the
position acquisition mode and its central wavelength is set to
zero. The image on the CCD, which is most likely blurred initially,
may be skeletonized by a computer algorithm to generate a thin
line. This thin line may represent the perfect image when the line
focusing is achieved. The third light may be directed by the mirror
728, the SLM 736, the DF 716 (and passing through LP 740) to
detector 706. There may be two alternative methods to implement
autofocusing.
[0134] In the first method, the root mean square deviation between
the line image and the skeletonized image may be used to guide the
spatial light modulator (SLM) 736 to shape the wavefront of light,
which may in turn change the focus of each point on the line until
the deviation between the two images (i.e. the line image and the
skeletonized image) is minimized. All points along the line focus
may be considered simultaneously.
[0135] The device or setup may include a spatial light modulator
feedback circuit coupling the detector 706 to the spatial light
modulator (SLM) 736. The spatial light modulator feedback circuit
may be configured to generate a skeletonized line (e.g. an initial
skeletonized line) based on a line (e.g. an initial line), formed
by the second light the sample 714 or organ). The spatial light
modulator feedback circuit may be configured to determine a root
mean square deviation between a subsequent line (i.e. the line
image) and the initial skeletonized line. The spatial light
modulator (SLM) 736 may be to adjust, based on a feedback from the
spatial light modulator feedback circuit, until the root mean
square deviation of the subsequent line (i.e. the line image) and
the initial skeletonized line is at a minimum.
[0136] In the second method, the focus index of each point in the
skeletonized image may be calculated and fed back to the spatial
light modulator (SLM) 736. The corresponding pixels in the spatial
light modulator (SLM) 736 may adjust their phase values to maximize
the focus index. Every point along the line focus may be considered
separately.
[0137] The device or setup may include a spatial light modulator
feedback circuit coupling the detector 706 to the spatial light
modulator (SLM) 736. The spatial light modulator feedback circuit
may be configured to generate a skeletonized line, e.g. an
skeletonized line based on a line, e.g. an initial line, formed by
the second light (on the sample 714 or organ). The (initial)
skeletonized line may be used as a reference line and a focus index
of each pixel along the (initial) skeletonized line may be used as
reference focus index. The spatial light modulator (SLM) 736 may be
configured to adjust, based a feedback from the spatial light
modulator feedback circuit, until a focus index of each pixel along
a subsequent skeletonized line generated reaches a maximum value.
The focus index of each pixel may be compared with the focus index
of a corresponding pixel of the other skeletonized lines, e.g. the
(initial) skeletonized line and the subsequent skeletonized lines.
A pixel with the highest value may be determined and the spatial
light modulator (SLM) 736 may be configured to be adjusted (by
adjusting the phase value of each pixel of the SLM 736) so that an
optimized skeletonized line including a plurality of pixels, each
pixel having a highest value compared to corresponding pixels of
other skeletonized line, may be generated on the sample 714.
[0138] Raman spectroscopy has been widely used in biomedical and
clinical applications. Raman spectroscopy measures the inelastic
scattering of photons induced by interaction with molecular bonds,
and may thus contain rich biochemical information. However, due to
inherently weak Raman signal, Raman data acquisition may be
generally slow, which may prohibit Raman spectroscopic imaging from
being used to investigate fast changing phenomena especially in
biological samples.
[0139] Several Raman imaging techniques may be developed to
overcome this limitation. Line scanned Raman imaging may collect
both spatial and spectral information along a line simultaneously.
Laser light may be shaped into a line using cylindrical optics or a
scanning mechanism and Raman spectra may be collected by a
two-dimensional detector array, i.e. charge-coupled device (CCD).
While spatial information is acquired along the laser line, the
spectral information may be collected in the dimension
perpendicular to the laser line. Although the data acquisition
speed is improved significantly, this method may cause
field-curvature artifacts. Its actual speed may be limited by the
requirement of autofocusing prior to data acquisition. Another
approach for Raman imaging may be based on acousto-optic tunable
filter (AOTF) or on liquid crystal tunable filter (LCTF). This
approach may benefit from the capability of transmitting a
selectable wavelength of light using tunable filters. However, the
disadvantage of this method may include long data acquisition when
the required spectral resolution is high. Fiber array Raman imaging
may also be a technique which could speed up Raman acquisition
significantly. Both spatial and spectral information may be
collected at the same time with a fiber array by rearranging
two-dimensional optical fibers array at the sample end to
one-dimensional array on the detector end. However, the number of
pixels in the fiber array may be limited by the amount of fibers
that could be mapped onto the CCD and the spatial resolution is
limited by the fiber size.
[0140] Reconstruction of the Raman signal from a few narrow-band
measurements at each pixel may realize the fast Raman imaging. Only
a few narrow-band Raman images may be required and the full Raman
spectrum at each pixel may be reconstructed. Due to the small
number of images required, the traditional spectral imaging setup,
i.e. using multiple filters in front of a CCD, may work well and
high spatial resolution as well as high spectral resolution may
also be achieved at the same time. Data acquisition may be much
faster than most current Raman imaging setups based on laser
scanning.
[0141] Various embodiments may include reconstructing Raman spectra
in the absence of fluorescence background. However, the drawback of
this method is that it may require a calibration data set. The
calibration data set may have to be similar to those obtained from
test samples. Therefore, when a new type of samples is tested, a
new calibration data set may be required and the size of the new
calibration data may be several dozens or even several hundreds.
Various embodiments may seek to find a method to suppress the
calibration data size or to reduce the need of measuring the
calibration data set for every new type of samples. Fluorescence
background may be present in a great variety of situations and its
magnitude may be often significant compared to the Raman signal
unless sophisticated techniques, such as shifted excitation Raman
difference spectroscopy, Fourier transformed Raman spectroscopy or
temporal gating, are employed to suppress fluorescence.
[0142] Various embodiments may involve measuring the basic
components, from which most test samples are made, instead of
similar samples for the calibration purpose. Every type of samples
may include several basic components, and the number of basic
components may be much smaller than the size of the traditional
calibration data set. In addition, different types of samples may
share same basic components. If those basic components have been
measured before, the repeated measurements of calibration data may
be reduced or eliminated.
[0143] The feasibility of using basic components as the calibration
data set has been demonstrated on 25 agar phantoms. The results
show the potential of using basic components instead of the
traditional calibration data set. This method may be extended to
cell spectra of an organ such as the eye.
[0144] The phantoms were made by mixing urea (V3171, Promega
corporation, US) and potassium formate (294454-500G, Sigma-Aldrich,
US) in 1.5% agar (PC0701-500G, Vivantis Technologies, US) dissolved
in distilled water. The concentrations of two calibration phantoms
were 1 M urea and 1 M potassium formate respectively. The
concentrations of 25 test phantoms for both urea and potassium
formate under investigation included 0.25 M, 0.5 M, 1 M, 1.5 M and
2 M. The two calibration phantoms were used as the calibration data
set and the 25 test phantoms were used as the test data set in this
study.
[0145] Raman spectra were measured over a range from 600 cm.sup.-1
to 1800 cm.sup.-1, by using a micro-Raman system (innoRam-7855,
B&W TEK, US) coupled to a video microscope sampling system
(BAC151A, B&W TEK, US). The excitation wavelength was 785 nm
and the spectral resolution was 4 cm.sup.-1. The exposure time for
Raman spectra was 10 s and each spectrum was accumulated for 30
times.
[0146] The narrow-band measurement c was simulated according to
Equation (3).
c=Fs+e (3)
where s (m.times.1 matrix, in which m is the number of wavenumbers)
is the Raman spectrum with fluorescence background, F (n.times.m
matrix, in which n is the number of filters) represents the
transmission spectra of the filters and e (n.times.1 matrix) is the
noise in narrow-narrow-band measurements. In Wiener estimation, a
Wiener matrix W (n.times.m matrix) is used to transform narrow-band
measurements c (m.times.1 matrix) into the corresponding Raman
spectrum s (n.times.1 matrix),
s =Wc (4)
so that the mean square error between the original and estimated
spectra is minimized. The Wiener matrix W is given by Equation
(4).
W=K.sub.sF.sup.T(FK.sub.sF.sup.T+K.sub.e).sup.-1 (5)
where
K.sub.s=E{ss.sup.T}, K.sub.e=E{ee.sup.T} (6)
In Equations (5) and (6), the superscript "T" represents matrix
transpose, the superscript "-1" represents matrix inverse and E{ }
represents an ensemble average. Plugging Equation (6) into Equation
(7) and ignoring the noise term yields
W=E{sc.sup.T}[E{cc.sup.T}].sup.-1 (7)
The PCs (principal components) based filter was used to generate
the narrow-band measurements. The relative root mean square error
(RMSE) of the reconstructed Raman spectrum after the removal of
fluorescence background, relative to the measured Raman spectrum in
which fluorescence background was also removed in the same manner,
was computed as in Eq. (8).
Relative RMSE = [ i = 1 N [ R r ( .lamda. i ) - R m ( .lamda. i ) ]
2 N .times. max [ R m ( .lamda. i ) ] 2 ] 1 / 2 ( 8 )
##EQU00003##
where R.sub.r and R.sub.m are the reconstructed Raman spectrum and
the measured Raman spectrum (both after fluorescence background
removed), respectively, .lamda..sub.i is the i-th wavenumber (i is
varied from 1 to N) and the function, max[ ], returns the maximum
intensity of the input spectrum.
[0147] Table 1 shows the relative RMSE for different number of
Principal Components (PCs) based filters. Because there're only two
basic components, the number of the PCs based filters was selected
up to 2. From the result, the relative RMSE improves significantly
when 2 PCs based filters were used. Table 1 shows the relative RMSE
for different number of PCs based filters
TABLE-US-00001 TABLE 1 PC number 1 2 Relative RMSE 13.9458
0.0642
[0148] FIG. 8A is a plot 800a of intensity (arbitrary units)
against wavenumber (cm.sup.-1) illustrating the best case of the
test phantoms. FIG. 8B is a plot 800b of intensity (arbitrary
units) against wavenumber (cm .sup.-1) illustrating a typical case
of the test phantoms. FIG. 8C is a plot 800c of intensity
(arbitrary units) against wavenumber (cm.sup.-1) illustrating the
worst case of the test phantoms. 802a represents the original Raman
signal while 802b represents the reconstructed Raman signal of the
best case. 802c represents the original Raman signal while 802d
represents the reconstructed Raman signal of the typical case. 802e
represents the original Raman signal while 802f represents the
reconstructed Raman signal of the worst case.
[0149] The relative RMSEs for the best, typical and worst case are
0.0135, 0.0703 and 0.2449, respectively. From the typical case and
the worst case, it may be noted that the peak intensities of the
reconstructed Raman signal are different from the peak intensities
of the original Raman signal, but the relative intensities between
the peaks within each sample is fairly close. The reconstructed
Raman spectrum and original Raman spectrum were then normalized by
dividing the intensity at each wavenumber by the sum of the
intensities at all wavenumbers. The relative RMSE improves
significantly to 0.0369 for 2 PCs based filters. As only one
concentration for each basic component is used, the mismatching
between the reconstructed and origin intensities may be due to the
lack of intensity information for different concentrations of basic
components. Various embodiments may include multiple concentrations
of the basic components for the calibration data set. The size of
the calibration data set may still be small because the large
amount of combinations of the basic components is avoided and the
need for the repeated measurements of the same basic components for
different types of samples may be removed or reduced.
[0150] Various embodiments may provide a spectral reconstruction
method based on Wiener estimation to reconstruct Raman spectra with
high spectral resolution from the narrow-band Raman measurements
with fluorescence background.
[0151] In another experiment, a genetic algorithm is used to
identify the optimal combination of different numbers of Gaussian
filters and commercial filters for spectral reconstruction to
improve accuracy. The importance of spectral information in the
Raman signal and that in fluorescence background were studied by
exploring two sets of components (PCs) based filters, derived
separately from principal component analysis (PCA) of clean Raman
signal and fluorescence background underlying it, for spectral
reconstruction. The new strategy was evaluated on both spontaneous
Raman data and SERS data, in which the former data represented the
case of high fluorescence while the latter data represented the
case of low fluorescence background. The results suggest the high
feasibility of eliminating the requirement of sophisticated Raman
system for fluorescence suppression in the reconstruction of Raman
spectra using the Wiener estimation based method. Various
embodiments may provide a method applicable to a simple and
inexpensive Raman setup for fast Raman imaging that involve most
Raman spectroscopy based applications.
[0152] Spontaneous Raman data were collected from live, apoptotic
and necrotic leukemia cells using a micro-Raman system (inVia,
Renishaw, UK) coupled to a microscope (Alpha 300, WITec, Germany)
in a backscattering setup. Ten Raman spectra from each group were
collected over a range from 600 to 1800 cm.sup.-1. The excitation
wavelength was 785 nm and the spectral resolution was 2
cm.sup.-1.
[0153] Surface enhanced Raman spectroscopy (SERS) data were
measured from blood serum samples collected from 50 patients with
nasopharyngeal cancer in Fujian Tumor hospital, Fuzhou, Fujian
Province, China. Blood serum samples were obtained by
centrifugation at 2,000 rpm for 15 minutes in order to remove blood
cells and then mixed with silver colloidal nanoparticles at a size
of 34 nm. The mixture was incubated for two hours at 4.degree. C.
before measurement. A confocal Raman micro-spectrometer (inVia,
Renishaw, UK) was used to measure Raman spectra over a range from
600 to 1800 cm.sup.-1 from human blood serum. The excitation
wavelength was 785 nm and the spectral resolution was 2 cm.sup.-1.
The details of sample preparation have been described in Feng et
al. (Biosensors and Bioelectronics, 26, 3167, 2011) and Lin et al.
Journal of Raman Spectroscopy, 43, 497, 2012).
[0154] The simulation of narrow-band measurements and methods of
reconstruction and evaluation were similar to those in the previous
study, which are briefly reiterated below. Since a filter is fully
characterized by its transmission spectrum, it may be reasonable to
expect that the result shown here faithfully mimics the real
situation in which Raman spectra are acquired by using these
filters. The narrow-band measurement c was simulated according to
Equation (9).
c=Fs (9)
where s (m.times.1 matrix, in which m is the number of wavenumbers)
is the Raman spectrum with fluorescence background, F (n.times.m
matrix, in which n is the number of filters) represents the
transmission spectra of the filters.
[0155] FIG. 9 is a schematic illustrating a procedure for Wiener
estimation according to various embodiments. Wiener estimation may
be used to reconstruct Raman spectra from simulated narrow-band
measurements, which was performed in two stages as shown in FIG. 9.
In the calibration stage, Wiener matrix was constructed, which
relates narrow-band measurements to the original Raman spectra
measured from samples in the calibration set. In the test stage,
Wiener matrix was applied to narrow-band measurements from an
unknown sample to reconstruct its Raman spectrum. The Wiener matrix
W may be defined in Equation (10), in which the noise term is
ignored.
W=E(sc.sup.T)[E(cc.sup.T)].sup.-1 (10)
where E( ) denotes the ensemble average, the superscript "T"
denotes matrix transpose and the superscript "-1" denotes matrix
inverse.
[0156] In various embodiments, the device may include a processor
coupled to the detector. The device may further include one or more
filters configured to generate one or more narrow-band Raman images
from an image (of the organ, e.g. the eye) captured by the
detector. The processor may be configured to generate a
reconstructed Raman image based on the one or more narrow Raman
images.
[0157] The one or more filters may be configured to generate one or
more reference narrow-band Raman images from a reference image. The
processor may be configured to determine a Wiener matrix based on
the one or more reference narrow-band Raman and the reference
image. The processor may be configured to generate the
reconstructed Raman image based on the one or more narrow Raman
images and the Wiener matrix.
[0158] Modified Wiener estimation, which is based on traditional
Wiener estimation, may improve reconstruction accuracy by
synthesizing new narrow-band measurements with an additional set of
filters. In the calibration stage, the modified Wiener matrix may
be computed by the combination of original narrow-band measurements
and the synthesized narrow-band measurements. In addition, two
strategies may be used to find the correction relations for
synthesizing new narrow-band measurements. In the test stage, new
narrow-band measurements were synthesized and corrected by the
correction relations obtained in the calibration stage. The
modified Wiener matrix was then applied and Raman spectra may be
reconstructed accurately because the new synthesized narrow-band
measurements may provide additional information. A final selection
step may be needed to select a better one from the results of
reconstructed generated using two correction relations. More
details about modified Wiener estimation have been described in
Chen et al (Journal of Biomedical Optics, 17, 0305011, 2012).
[0159] In order to evaluate the accuracy of a reconstructed Raman
spectrum, the reconstructed Raman spectrum may be first
preprocessed to remove fluorescence background by using the
fifth-order polynomial fitting. Then the relative root mean square
error (RMSE) of the reconstructed Raman spectrum after the removal
of fluorescence background, relative to the measured Raman spectrum
in which fluorescence background may also removed in the same
manner, was computed as in Equation (8).
[0160] Four different categories of filters were examined in this
experiment, which include commercial filters, Gaussian filters,
principal components (PCs) based filters and non-negative PCs based
filters. These commercial filters, Gaussian filters, PCs based
filters and non-negative PCs based filters described herein are
examples and are not intended to be limiting. Table 2 illustrates
the commercial filters used in the simulations of narrow-band
measurements.
TABLE-US-00002 TABLE 2 Manufacturer Product numbers of commercial
filters Chroma D850/20m, D850/40m Technique (Bellows Falls, VT, US)
Edmund Optics NT 84-790, NT 84-791 (Barrington, NJ, US) Omega
Filters 3RD850LP, 3RD900LP, XB 142, XB 143, XB 146, XB
(Brattleboro, 149, XF 3308, XL 19, XL 40, XLK 18, XLK 20 VT, US)
Semrock FF 01-830/2-25, FF 01-832/37-25, FF 01-835/70-25, FF
(Rochester, 01-840/12-25, FF 01-857/30-25, FF 01-910/5-25 NY, US)
Thorlabs FB 830-10, FB 840-10, FB 850-10, FB 850-40, FB (Newton,
860-10, FB 870-10, FB 880-10, FB 880-40, FB 890-10, NJ, US) FB
900-40, FB 910-10, FL 830-10, FL 850-10, FL 880-10, FL 905-10, FL
905-25
[0161] A total of 37 commercial filters from five manufacturers
were investigated as shown in Table 2. The transmittance spectra of
these filters at least partially overlap with the range of about
600 to 1 about 800 cm.sup.-1 an excitation wavelength of 785
nm.
[0162] A collection of 72 Gaussian filters were synthesized
numerically in this study. A Gaussian filter may be expressed
mathematically as
G ( .lamda. ) = exp ( - ( .lamda. - u ) 2 2 .sigma. 2 ) ( 11 )
##EQU00004##
where G(A) denotes the transmittance at the wavelength A,
represents the central wavelength and a represents the standard
deviation. The central wavelength was varied over a range from 830
nm to 910 nm and the increment was 10 nm. The standard deviation
was varied over a range from 2.5 nm to 20 nm and the increment was
2.5 nm.
[0163] Both PCs based filters and non-negative PCs based filters
were derived using the principle component analysis (PCA) method.
In this experiment, the transmittance spectra of PCs based filters
were equivalent in shape to the first several PCs of the Raman
spectra with fluorescence background. The transmittance spectra of
non-negative PCs based filters were generated using the same method
as in Piche (Journal Optical Society of America A, 19, 1946,
2002).
[0164] Genetic algorithm may be usually used to generate useful
solutions for optimization and search problems, which is based on
the evolution, i.e. the survival of the fittest strategy. In the
experiment, genetic algorithm has been used to find the optimal
combination of Gaussian filters and that of commercial filters to
achieve a minimal RMSE in reconstructed Raman spectra. The
optimization methodology proceeded in the following manner.
Firstly, a population of filter combination was initialized
randomly. Secondly, Wiener estimation was applied to reconstruct
Raman spectra and the mean accuracy of the reconstructed Raman
spectra was evaluated. Thirdly, a new population of filter
combination was generated according to the mean accuracy of the
reconstructed Raman spectra, in which the filter combination
yielding higher reconstruction accuracy is be more likely to become
the parent for the generation of the new population. The crossover
rate was 0.9 and the mutation rate was 0.1. The second and third
steps were repeated iteratively until an optimized combination of
filters was found. The optimization method was coded and run in
Matlab (MATLAB R2011b, MathWorks, Natick, Mass., US).
[0165] The leave-one-out method was used for cross-validation in
the experiment to fully utilize each sample in an unbiased manner.
The measurement from one sample was used as the test data each time
and the measurements from the rest of samples were used as the
calibration data set. This procedure was repeated until the
measurement from every sample has been tested once. For Gaussian
and commercial filters, a new set of the optimal filters and Wiener
matrix were generated from the calibration data set by the genetic
algorithm in each round of the cross-validation and then applied to
the test data. For PCs based filters and non-negative PCs based
filters, the filters were fixed, thus it was not necessary to find
the optimal filters. Only Wiener matrix was generated from the
calibration data set in each round and then applied to the test
data.
[0166] FIG. 10A is a plot of intensity (arbitrary units) against
wavenumber (cm.sup.-1) illustrating an original spontaneous Raman
data, including fluorescence background, of leukemia cells. FIG.
10B is a plot of intensity (arbitrary units) against wavenumber
(cm.sup.-1) illustrating an original surface enhanced Raman
spectroscopy (SERS) data, including fluorescence background, of
blood serum sample. Compared with SERS spectra from blood serum
sample in FIG. 10B, the spontaneous Raman data from leukemia cells
in FIG. 10A show higher variance in fluorescence background, which
may be shown to affect the accuracy of reconstructed spectra.
[0167] Table 3 shows the cumulative contribution ratio of different
PC numbers for spontaneous Raman spectra and SERS spectra. The
cumulative contribution ratio refers to the ratio of the sum of
eigenvalues corresponding to PCs of interest to the sum of all
eigenvalues. By using up to six filters, a high percentage of
99.99% is reached in both sets of Raman spectra. Therefore, we test
the filter number from three to six, which should be sufficient for
Raman reconstruction with high accuracy.
[0168] Table 3 compares the cumulative contribution ratio of
different PC numbers for spontaneous Raman spectra and SERS
spectra.
TABLE-US-00003 TABLE 3 Spontaneous SERS PC number Raman spectra (%)
spectra (%) 2 99.69 99.96 3 99.89 99.98 4 99.95 99.99 5 99.99 99.99
6 99.99 99.99 7 99.99 100.00
[0169] Table 3 shows the cumulative contribution ratio of different
PC numbers for spontaneous Raman spectra and SERS spectra. The
cumulative contribution ratio refers to the ratio of the sum of
eigenvalues corresponding to PCs of interest to the sum of all
eigenvalues. By using up to six filters, a high percentage of
99.99% is reached in both sets of Raman spectra.
[0170] Three to six filters have been tested, which should be
sufficient for Raman reconstruction with high accuracy. Table 4
compares the mean relative RMSE of spontaneous Raman spectra (after
fluorescence background removed) reconstructed from narrow-band
measurements using different types and numbers of filters.
TABLE-US-00004 TABLE 4 Non-negative Commercial Gaussian PCs based
PCs based filters filters filters filters 3 filters 5.61 .times.
10.sup.-2 5.18 .times. 10.sup.-2 6.93 .times. 10.sup.-2 6.93
.times. 10.sup.-2 4 filters 4.91 .times. 10.sup.-2 5.07 .times.
10.sup.-2 5.83 .times. 10.sup.-2 5.83 .times. 10.sup.-2 5 filters
4.01 .times. 10.sup.-2 4.37 .times. 10.sup.-2 3.20 .times.
10.sup.-2 3.20 .times. 10.sup.-2 6 filters 3.49 .times. 10.sup.-2
3.54 .times. 10.sup.-2 2.57 .times. 10.sup.-2 2.57 .times.
10.sup.-2
[0171] Table 4 shows the comparison in the mean relative RMSE of
spontaneous Raman spectra (after fluorescence background removed)
reconstructed from narrow-band measurements using different types
and numbers of filters. The percentage values of reduction in the
mean relative RMSE from three to four filters were 12.5%, 2.1%,
15.9% and 15.9% for commercial filters, Gaussian filters, PCs based
filters and non-negative PCs based filters, respectively. The
percentage values of reduction from four to five filters were
18.3%, 13.8%, 45.1% and 45.1% and the reduction from five to six
filters were 13.0%, 19.0%, 19.7% and 19.7%, respectively.
[0172] FIG. 11A is a plot 1100a of intensity (arbitrary units)
against wavenumber (cm.sup.-1) illustrating measured spontaneous
Raman spectrum and the spontaneous Raman spectrum reconstructed by
traditional Wiener estimation for the best case using the best
combination of six commercial filters. 1102a indicates the measured
spontaneous Raman spectrum while 1102b indicates the reconstructed
spectrum. FIG. 11B is a plot 1100b of intensity (arbitrary units)
against wavenumber (cm.sup.-1) illustrating measured spontaneous
Raman spectrum and the spontaneous Raman spectrum reconstructed by
traditional Wiener estimation for a typical case using the best
combination of six commercial filters. 1102c indicates the measured
spontaneous Raman spectrum while 1102d indicates the reconstructed
spectrum. FIG. 11C is a plot 1100c of intensity (arbitrary units)
against wavenumber (cm.sup.-1) illustrating measured spontaneous
Raman spectrum and the spontaneous Raman spectrum reconstructed by
traditional Wiener estimation for the case using the best
combination of six commercial filters. 1102e indicates the measured
spontaneous Raman spectrum while 1102f indicates the reconstructed
spectrum. FIG. is a plot 1100d of intensity (arbitrary units)
against wavenumber (cm.sup.-1) illustrating the transmittance
spectra of the six commercial filters corresponding to the typical
case. 1104a represents the transmittance spectra of FB 860-10,
1104b represents the transmittance spectra of NT 84-791, 1104c
represents the transmittance spectra of FB 900-40, 1104d represents
the transmittance spectra of FF 01-857, 1104e represents the
transmittance spectra of XLK 20 and 1104f represents the
transmittance spectra of XB 143. The fluorescence background has
been removed in both sets of spectra (measured spontaneous Raman
spectrum and the spontaneous Raman spectrum reconstructed by
traditional Wiener estimation) to facilitate comparison in Raman
features.
[0173] FIGS. 11A-C show the comparison between the measured
spontaneous Raman spectra and the spontaneous Raman spectra
reconstructed by traditional Wiener estimation. FIG. 11D shows the
transmittance spectra of six commercial filters corresponding to
the typical case, i.e. FB 860-10, NT 84-791, FB 900-40, FF 01-857,
XLK 20 and XB 143. The fluorescence background has been removed in
both sets of spectra, i.e. the measured spontaneous Raman spectra
and the spontaneous Raman spectra reconstructed by traditional
Wiener estimation, to facilitate comparison in Raman features. The
typical case (shown in FIG. 11B) is the reconstructed spontaneous
Raman spectrum with a relative RMSE close to the mean relative
RMSE, while the best case (shown in FIG. 11A) and worst case (shown
in FIG. 11C) are the reconstructed spontaneous Raman spectra with
the minimum relative RMSE and maximum relative RMSE. The relative
RMSEs are 1.57.times.10.sup.-2, 3.29.times.10.sup.-2,
6.69.times.10.sup.-2 in the best case, the typical case and the
worst case, respectively.
[0174] FIG. 12A is a plot 1200a of intensity (arbitrary units)
against wavenumber (cm.sup.-1) illustrating measured spontaneous
Raman spectrum and the spontaneous Raman spectrum reconstructed by
traditional Wiener estimation for the best case using the best
combination of six non-negative principal components (PCs) based
filters. 1202a indicates the measured spontaneous Raman spectrum
while 1202b indicates the reconstructed spectrum. FIG. 12B is a
plot 1200b of intensity (arbitrary units) against wavenumber
(cm.sup.-1) illustrating measured spontaneous Raman spectrum and
the spontaneous Raman spectrum reconstructed by traditional Wiener
estimation for a typical case using the best combination of six
non-negative principal components (PCs) based filters. 1202c
indicates the measured spontaneous Raman spectrum while 1202d
indicates the reconstructed spectrum. FIG. 12C is a plot 1200c of
intensity (arbitrary units) against wavenumber (cm.sup.-1)
illustrating measured spontaneous Raman spectrum and the
spontaneous Raman spectrum reconstructed by traditional Wiener
estimation for the case using the best combination of six
non-negative principal components (PCs) based filters. 1202e
indicates the measured spontaneous Raman spectrum while 1202f
indicates the reconstructed spectrum. FIG. 12D is a plot 1200d of
intensity (arbitrary units) against wavenumber (cm.sup.-1)
illustrating the transmittance spectra of the six non-negative
principal components (PCs) based filters corresponding to the
typical case. 1204a represents the transmittance spectra of the
first filter, 1204b represents the transmittance spectra of the
second filter, 1204c represents the transmittance spectra of the
third filter, 1204d represents the transmittance spectra of the
fourth filter, 1204e represents the transmittance spectra of the
fifth filter and 1204f represents the transmittance spectra of the
sixth filter. The fluorescence background has been removed in both
sets of spectra (measured spontaneous Raman spectrum and the
spontaneous Raman spectrum reconstructed by traditional Wiener
estimation) to facilitate comparison in Raman features.
[0175] FIG. 12A-C show the comparison between the measured
spontaneous Raman spectra and the spontaneous Raman spectra
reconstructed by traditional Wiener with the first six non-negative
PCs based filters. The relative RMSEs were 9.3.times.10.sup.-3,
.times.10.sup.-2, 4.99.times.10.sup.-2 in the best case, the
typical case and the worst case, respectively.
[0176] Table 5 shows the comparison in the mean relative RMSE of
reconstructed SERS spectra (after fluorescence background removed)
from narrow-band measurements using different types and numbers of
filters. The percentage values of reduction in the mean relative
RMSE from three to four filters were 2.7%, 2.7%, 5.1% and 5.1% for
commercial filters, Gaussian filters, PCs based filters and
non-negative PCs based filters, respectively. The percentage values
of reduction from four to five filters were 1.2%, 3.9%, 19.3% and
19.3% and the percentage values of reduction from five to six
filters were 15.0%, 10.1%, 11.7% and 11.7%, respectively.
[0177] Table 5 compares the mean relative RMSE of SERS spectra
(after fluorescence background removed) reconstructed from
narrow-band measurements using different types and numbers of
filters.
TABLE-US-00005 TABLE 5 Non-negative Commercial Gaussian PCs based
PCs based filters filters filters filters 3 filters 2.65 .times.
10.sup.-2 2.64 .times. 10.sup.-2 2.13 .times. 10.sup.-2 2.13
.times. 10.sup.-2 4 filters 2.57 .times. 10.sup.-2 2.57 .times.
10.sup.-2 2.02 .times. 10.sup.-2 2.02 .times. 10.sup.-2 5 filters
2.54 .times. 10.sup.-2 2.47 .times. 10.sup.-2 1.63 .times.
10.sup.-2 1.63 .times. 10.sup.-2 6 filters 2.16 .times. 10.sup.-2
2.22 .times. 10.sup.-2 1.44 .times. 10.sup.-2 1.44 .times.
10.sup.-2
[0178] FIG. 13A is a plot 1300a of intensity (arbitrary units)
against wavenumber (cm.sup.-1) illustrating measured surface
enhanced Raman spectroscopy (SERS) spectrum and the surface
enhanced Raman spectroscopy (SERS) Raman spectrum reconstructed by
traditional Wiener estimation for the best case using the best
combination of six commercial filters. 1302a indicates the measured
surface enhanced Raman spectroscopy (SERS) spectrum while 1302b
indicates the reconstructed spectrum. FIG. 13B is a plot 1300b of
intensity (arbitrary units) against wavenumber (cm.sup.-1)
illustrating measured surface enhanced Raman spectroscopy (SERS)
spectrum and the surface enhanced spectroscopy (SERS) spectrum
reconstructed by traditional Wiener estimation for a case using the
best combination of six commercial filters. 1302c indicates the
measured surface enhanced Raman spectroscopy (SERS) spectrum while
1302d indicates the reconstructed spectrum. FIG. 13C is a plot
1300c of intensity (arbitrary units) against wavenumber (cm.sup.-1)
illustrating measured surface enhanced Raman spectroscopy (SERS)
spectrum and the surface enhanced Raman spectroscopy (SERS)
spectrum reconstructed by traditional Wiener estimation for the
worst case using the best combination of six commercial filters.
1302e indicates the measured surface enhanced Raman spectroscopy
(SERS) spectrum while 1302f indicates the reconstructed spectrum.
FIG. 13D is a plot 1300d of intensity (arbitrary units) against
wavenumber (cm.sup.-1) illustrating the transmittance spectra of
the six commercial filters corresponding to the typical case .
1304a represents the transmittance spectra of FB 840-10, 1304b
represents the transmittance spectra of FF 01-857, 1304c represents
the transmittance spectra of FL 905-10, 1304d represents the
transmittance spectra of FB 850-40, 1304e represents the
transmittance spectra of FF01-840 and 1304f represents the
transmittance spectra of XB 149. The fluorescence background has
been removed in both sets of spectra (measured surface enhanced
Raman spectroscopy (SERS) spectrum and the surface enhanced
spectroscopy (SERS) spectrum reconstructed by traditional Wiener
estimation) to facilitate comparison in Raman features.
[0179] FIGS. 13A-C show the comparison between the measured SERS
spectrum and the SERS spectrum reconstructed by traditional Wiener
estimation and the transmittance spectra of the best combination of
six commercial filters corresponding to the typical case, i.e. FL
840-10, FF 01-857, FL 905-10, FB 850-40, FF 01-840 and XB 149. The
relative RMSEs were 9.1.times.10.sup.-3, 2.13.times.10.sup.-2 and
5.76.times.10.sup.-2 in the best case, the typical case and the
worst case, respectively.
[0180] FIG. 14A is a plot 1400a of intensity (arbitrary units)
against wavenumber (cm.sup.-1) illustrating measured surface
enhanced Raman spectroscopy (SERS) spectrum and the surface
enhanced Raman spectroscopy (SERS) spectrum reconstructed by
traditional Wiener estimation for the best case using the best
combination of six non-negative principal components (PCs) based
filters. 1402a indicates the measured surface enhanced Raman
spectroscopy (SERS) spectrum while 1402b indicates the
reconstructed spectrum. FIG. 14B is a plot 1400b of intensity
(arbitrary units) against wavenumber (cm.sup.-1) illustrating
measured surface enhanced Raman spectroscopy (SERS) spectrum and
the surface enhanced Raman spectroscopy (SERS) spectrum
reconstructed by traditional Wiener estimation for a typical case
using the best combination of six non-negative principal components
(PCs) based filters. 1402c indicates the measured surface enhanced
Raman spectroscopy (SERS) spectrum while 1402d indicates the
reconstructed spectrum. FIG. 14C is a plot 1400c of intensity
(arbitrary units) against wavenumber (cm.sup.-1) illustrating
measured surface enhanced Raman spectroscopy (SERS) spectrum and
the surface enhanced Raman spectroscopy (SERS) spectrum
reconstructed by traditional Wiener estimation for the worst case
using the best combination of six non-negative principal components
(PCs) based filters. 1402e indicates the measured surface enhanced
Raman spectroscopy (SERS) spectrum while 1402f indicates the
reconstructed spectrum. FIG. 14D is a plot 1400d of intensity
(arbitrary units) against wavenumber (cm.sup.-1) illustrating the
transmittance spectra of the six non-negative principal components
(PCs) based filters corresponding to the typical case. 1404a
represents the transmittance spectra of the first filter, 1404b
represents the transmittance spectra of the second filter, 1404c
represents the transmittance spectra of the third filter, 1404d
represents the transmittance spectra of the fourth filter, 1404e
represents the transmittance spectra of the fifth filter and 1404f
represents the transmittance spectra of the sixth filter. The
fluorescence background has been removed in both sets of spectra
(measured surface enhanced Raman spectroscopy (SERS) spectrum and
the surface enhanced Raman spectroscopy (SERS) spectrum
reconstructed by traditional Wiener estimation) to facilitate
comparison in Raman features.
[0181] FIGS. 14A-C show the comparison between the measured SERS
spectra and the SERS spectra reconstructed by traditional Wiener
estimation with the first six non-negative PCs based filters. The
relative RMSEs were 8.5.times.10.sup.-3, 1.44.times.10.sup.-2,
2.85.times.10.sup.-2 in the best case, the typical case and the
worst case, respectively.
[0182] It has been demonstrated that full Raman spectra may be
reconstructed by Wiener estimation from a few narrow-band
measurements in the presence of fluorescence background. The
experiment has proved the feasibility of applying reconstruction
based on a plurality of narrow-band measurements by Wiener
estimation, enabling fast Raman imaging using a simple Raman
setup.
[0183] For SERS spectra, Gaussian filters and commercial filters
showed worse accuracies compared with PCs based filters. This may
be attributed to the ability of PCs based filter to capture more
variance, i.e. information, compared with Gaussian filters and
commercial filters. In addition, the importance of capturing both
Raman signal and fluorescence background information have also been
verified. SERS spectra generated from the Raman signal or
fluorescence background alone using PCs based filters have been
tested. For SERS spectra generated from Raman signal using PCs
based filters, the relative RMSEs were 4.57.times.10.sup.-2,
4.47.times.10.sup.-2, 3.28.times.10.sup.-2 and 3.02.times.10.sup.-2
for three, four, five and six filters, respectively. For SERS
spectra generated from fluorescence background using PCs based
filters, the relative RMSEs were 2.40.times.10.sup.-2,
2.33.times.10.sup.-2, 2.14.times.10.sup.-2 and 2.03.times.10.sup.-2
for three, four, five and six filters, respectively. Both sets of
relative RMSEs were considerably greater than those obtained using
PCs based filters generated from Raman spectra with fluorescence
background as shown in Table 5, which implies that information from
both Raman signal and fluorescence background is important for
reconstruction.
[0184] For spontaneous Raman spectra, the mean values of
reconstruction accuracy using Gaussian filters and commercial
filters were better than PCs based filters when only three or four
filters were used. This may be attributed to the fluorescence
background much larger in spontaneous Raman spectra compared to
that in SERS spectra. In this case, the variance for fluorescence
background in a spontaneous Raman spectrum was considerably larger
than the Raman signal in magnitude. Based on the characteristics of
PCA, the first three or four PCs, from which the transmittance
spectra of these PCs based filters were derived, capture most
information from smooth fluorescence background and less
information from the Raman signal on top of the fluorescence
background. Interestingly, PCs based filters showed better
reconstruction accuracy than Gaussian filters and commercial
filters when five or six filters were used. Moreover, the
improvement in accuracy for PCs based filters from four to five
filters was significant, which means that more information about
Raman signal was collected by the additional PCs based filters as
sufficient information about fluorescence background has been
collected by the first four PCs based filters. Spontaneous Raman
spectra were also generated from the Raman signal or fluorescence
background alone using PCs based filters. For spontaneous Raman
spectra generated from the Raman signal using PCs based filters,
the relative RMSEs were 1.43.times.10.sup.-1, 1.43.times.10.sup.-1,
5.60.times.10.sup.-2 and 5.42.times.10.sup.-2 for three, four, five
and six filters, respectively. For spontaneous Raman spectra
generated from fluorescence background using PCs based filters, the
relative RMSEs were 8.86.times.10.sup.-2, 8.28.times.10.sup.-2,
9.69.times.10.sup.-2 and 8.12.times.10.sup.-2 for three, four, five
and six filters, respectively. These values were much larger than
those obtained using PCs based filters generated from Raman spectra
with fluorescence background as shown in Table 4. This further
shows the importance of deriving optimal filters from both Raman
signal and fluorescence background.
[0185] For both spontaneous Raman spectra and SERS spectra,
additional filters may improve the reconstruction accuracy
significantly. Therefore, a tradeoff between the accuracy and cost
needs to be made in the choice of number of filters. Compared with
spontaneous Raman spectra, the reconstruction accuracy of SERS
spectra was much better when the same number of filters were used
as shown in Tables 4 and 5. This observation could be explained by
two factors. One is that SERS spectra may contain smaller
fluorescence background than spontaneous Raman spectra, which
lowers down the requirement on the effectiveness of the filter set
in capturing most information. The other is that SERS spectra may
exhibit higher signal-to-noise ratio, which reduces the influence
of noise on reconstruction. Using sophisticated methods, e.g.
shifted excitation Raman difference spectroscopy, Fourier
transformed Raman spectroscopy, and temporal gating, to suppress
fluorescence background and/or improve the signal-to-noise ratio of
Raman signals would further improve the reconstruction
accuracy.
[0186] Table 6 compares the relative RSME of spontaneous Raman
spectra (after fluorescence background removed) reconstructed from
narrow-band measurements with the best combination of three filters
using traditional Wiener estimation and between modified Wiener
estimation.
TABLE-US-00006 TABLE 6 Non-negative Commercial Gaussian PCs based
PCs based filters filters filters filters Traditional 5.61 .times.
10.sup.-2 5.18 .times. 10.sup.-2 6.93 .times. 10.sup.-2 6.93
.times. 10.sup.-2 Wiener estimation Modified 4.82 .times. 10.sup.-2
4.55 .times. 10.sup.-2 6.99 .times. 10.sup.-2 7.13 .times.
10.sup.-2 Wiener estimation
[0187] Table 7 compares the relative RSME of SERS spectra (after
fluorescence background removed) reconstructed from narrow-band
measurements with the best combination of three filters using
traditional Wiener estimation and between modified Wiener
estimation.
TABLE-US-00007 TABLE 7 Non-negative Commercial Gaussian PCs based
PCs based filters filters filters filters Traditional 2.65 .times.
10.sup.-2 2.64 .times. 10.sup.-2 2.13 .times. 10.sup.-2 2.13
.times. 10.sup.-2 Wiener estimation Modified 2.54 .times. 10.sup.-2
2.61 .times. 10.sup.-2 2.14 .times. 10.sup.-2 2.15 .times.
10.sup.-2 Wiener estimation
[0188] In addition, the method of modified Wiener estimation
developed previously have been compared to traditional Wiener
estimation for both spontaneous Raman spectra and SERS spectra as
shown in Tables 6 and 7. For spontaneous Raman spectra shown in
Table , there were reduction in percentage values of 14.1% and
12.2% in the relative RMSE for commercial filters and Gaussian
filters when using modified Wiener estimation compared with the
traditional Wiener estimation. In contrast, there was small
degradation in reconstruction accuracy from traditional Wiener
estimation to modified Wiener estimation for PCs based filters and
non-negative PCs based filters.
[0189] These observations may be explained as below. In modified
Wiener estimation, although additional information was provided by
using synthesized narrow-narrow-band measurements, error was also
induced with the correction process. The final reconstruction
accuracy was a compromise between the gain (the additional
information) and the loss (the induced error). For commercial and
Gaussian filters, the additional filters created in the modified
Wiener estimation were the first three PCs based filters. In
contrast, the additional filters were the fourth to sixth PCs based
filters for PCs based filters and non-negative PCs based filters
because the first three PCs filters have been applied. The first
three PCs based filters with relatively smooth shapes were likely
to capture additional information (mainly from slow changing
fluorescence background), which outweighed the error induced
(mainly from sharp Raman signals) in the correction process. In
comparison, the fourth to sixth PCs based filters with sharp peaks
were likely to induce larger errors in the correction process
because it can only capture additional information mainly from
similarly sharp Raman peaks, which were more difficult to correct.
For SERS spectra, which contained much weaker fluorescence
background than the spontaneous Raman spectra, shown in Table 7,
there was no considerable difference in terms of the relative RMSE
between modified Wiener estimation and traditional Wiener
estimation. Therefore, the method of modified Wiener estimation may
show more significant advantage over traditional Wiener estimation
in Raman spectra with intense fluorescence background.
[0190] In practice, a new Wiener matrix may need to be constructed
with a new set of the calibration data when using a different type
of sample. While this may be the major limitation for this method,
this method may still find a large number of biomedical
applications, such as differentiation of cancer from normal samples
and classification of cell death mode etc. Most popular methods for
such applications rely on multi-variate statistical analysis, thus
also requiring a set of data for training the classifier.
[0191] This method of reconstruction may be advantageous when
employed in Raman imaging. Currently, most Raman imaging techniques
use point scanning or line scanning, in which every scan would
involve the acquisition of Raman intensity at many wavenumbers.
While wide-field Raman imaging using a CCD may be performed at each
wavenumber, this may require a filter with extremely narrow pass
band and tunable central wavelength. Moreover, it may be very time
consuming given the number of wavenumbers involved (hundreds to
thousands depending on spectral resolution required). Various
embodiments may require only a few narrow-band filters with much
larger bandwidths to get a few Raman images and the full Raman
spectrum at each pixel may be reconstructed. The potential
improvement in the speed may be dramatic just considering the
difference in the number of Raman images required between
traditional Raman imaging and the proposed strategy.
[0192] The experiment has demonstrated a spectral reconstruction
method based on Wiener estimation applied to the narrow-band Raman
measurements with fluorescence background for reconstructing the
Raman spectra with high spectral resolution. The reconstruction
method has been evaluated on both spontaneous Raman data and SERS
data. A genetic algorithm was used to identify the optimal
combination of different and types of filters for spectral
construction. The agreement between reconstructed spectra and
measured spectra was excellent in either set of data, which
indicates that this method may be applied to both spontaneous Raman
measurements and SERS measurements that involve most Raman
spectroscopy based applications. The reconstruction of SERS spectra
showed even better results, which demonstrates that the higher
signal to noise ratio and lower fluorescence background may improve
the reconstruction accuracy. For both spontaneous Raman spectra and
SERS spectra, the reconstruction accuracy may be improved
significantly by using additional filters and information from both
Raman signal and fluorescence background is important. to our
pervious study, the new results suggest that the proposed method
may be used in a simple Raman system that acquires Raman spectra
with fluorescence background. Therefore, this method may open a new
avenue for Raman imaging to investigate fast changing phenomena in
biomedical applications in a simple optical setup without the
function of fluorescence suppression.
[0193] FIG. 15 is a schematic 1500 illustrating a method of
operating a device for determining a condition of an organ of
either a human or an animal according to various embodiments. The
method may include, in 1502, activating a switching mechanism to
switch between an optical examination mode and a Raman mode. During
the optical examination mode, a lens system may be configured to
direct a first light emitted from a first optical source. During
the Raman mode, the lens system may be configured to direct a
second light emitted from a second optical source. Further, during
the Raman mode, the lens systems may be further configured to
direct a third light to the detector.
[0194] In other words, the method may include switching between an
optical examination mode and a Raman mode. The lens systems may be
configured to direct a first light emitted from a first optical
source during the optical examination mode. The lens system may be
configured to direct a second light emitted from a second optical
source during the Raman mode.
[0195] The first light emitted from the first optical source may
incident on an organ such as an eye. A fourth light may be
reflected from the organ when the first light is incident on the
organ. The fourth light may be derived from the first light.
[0196] The second light emitted from the second optical source may
incident on the organ. The third light may be reflected from the
organ when the second light is incident on the organ. The third
light may be derived from the second light. Raman analysis may be
based on the second light and third light.
[0197] A method of diagnosing an organ such as the eye may also be
provided. The method may include activating a switching mechanism
to switch to an optical examination mode so that a first light may
be emitted from a first optical source, such as a non-coherent
light source, to the organ and detecting a disease or anomaly of
the organ (by an observer such as a doctor or an optometrist) based
on a second light reflected from the organ. The method may further
include activating the switching mechanism to switch to a Raman
examination mode so that a second light may be emitted from a
second optical source and detecting (via a detector) a third light
reflected from the organ.
[0198] A method of operating a device may also be provided. The
method may activating a switching mechanism to switch between an
optical examination mode and a Raman mode. During the optical
examination mode, a lens system may be configured to direct a first
light emitted from a first optical source. During the Raman mode,
the lens system may be configured to direct a second light emitted
from a second optical source. Further, during the Raman mode, the
lens systems may be further configured to direct a third light to
the detector.
[0199] During the optical examination mode, a lens system may be
configured to direct the first light emitted from the first optical
source to an interface portion.
[0200] The lens system during the optical examination mode may be
further configured to direct a fourth light, the fourth light
derived from the first light, from the interface portion to an
optical examination output portion.
[0201] The lens system during the Raman mode may be configured to
direct the second light emitted from the second optical source to
the interface portion.
[0202] The lens system during the Raman mode may be configured to
direct the third light from the interface to a detector.
[0203] The third light may be derived from the second light. The
fourth light may be derived from the first light.
[0204] In various embodiments, a use of a device may be provided.
In various embodiments, a use of a device for determining a
condition of an organ of either a human may be provided. The method
may include activating a switching mechanism to switch between an
optical examination mode and a Raman mode. During the optical
examination mode, a lens system may be configured to direct a first
light emitted from a first optical source. During the Raman mode,
the lens system may be configured to direct a second light emitted
from a second optical source. Further, during the Raman mode, the
lens systems may be further configured to direct a third light to
the detector.
[0205] FIG. 16 is a schematic 1600 of a device for determining a
condition of an of either a human or an animal according to various
embodiments. The device may an optical source 1604, a detector 1606
and a lens system 1608. The lens system 1608 be configured to
direct a light emitted from the optical source 1604. The lens
system may be further configured to direct a further light to the
detector 1608. The further light may be based on or derived from
the light. The further light may be the light reflected by the
light incident on the organ. The lens system 1608 may be configured
to direct the to the organ. The further light reflected from the
organ may be directed by the lens 1608 to the detector 1606.
[0206] In various embodiments, the device illustrated in FIG. 1 may
only include the components for Raman mode. The components in the
device shown in FIG. 16 may work in a similar manner as the Raman
components in the device shown in FIG. 1.
[0207] The lens system 1608 may include an objective lens for
focusing the light. The lens system 1608 may further include an
actuator, e.g. a piezoelectric transducer, for controlling a
position of the objective lens. The lens system 1608 may include an
actuator feedback circuit coupling the detector to the
actuator.
[0208] The actuator feedback circuit may be configured to receive
an output from the detector 1606 and further configured to provide
a feedback to the actuator based on the output from the detector
1606. The actuator feedback circuit may be configured to determine
a focus index based on the output from the detector 1606 and
further configured to provide a feedback based on the determined
focus index and a reference focus index.
[0209] The lens system may include a spatial light modulator (or
another dynamic optical element such as a digital micromirror) for
modulating the light emitted from the optical source reflected. The
lens system may include a spatial light modulator feedback circuit
(or dynamic optical element feedback circuit) coupling the detector
1606 to the spatial light modulator or dynamic optical element. The
spatial light modulator feedback circuit (or dynamic optical
element feedback circuit) may be configured to generate a
skeletonized line based on a line formed by the light. The spatial
light modulator (or dynamic optical element feedback circuit) may
be configured to be adjusted based on a feedback from the spatial
light modulator feedback circuit (or dynamic optical element
feedback circuit) until a focus index of each pixel along a
subsequent skeletonized line generated reaches a maximum value. In
other words, the lens system may include a dynamic optical element
for modulating the light emitted from the optical source reflected
(to the organ). The dynamic optical element may be a spatial light
modulator (SLM) or a digital micromirror device. The lens system
may further include a dynamic optical element feedback circuit
coupling the detector to the dynamic optical element. The dynamic
optical element feedback circuit may be configured to generate a
skeletonized line based on a line formed by the light.
[0210] The lens system 1608 may include a single beam splitter or a
single dichroic mirror configured to direct the light (and/or
direct the further light). The device may further include a
processor coupled to the detector 1606.
[0211] The device may include one or more filters configured to
generate one or narrow-band Raman images from an image captured by
the detector 1606. The processor may be configured to generate one
or more reconstructed Raman images based on the one or more narrow
Raman images, each of the one or more reconstructed Raman images
corresponding to one wavelength. The processor may be further
configured to generate a Raman spectrum at each pixel based on the
one or more reconstructed Raman images. one or more filters may be
configured to generate one or more reference narrow-band Raman
images from one or more reference images that contain full spectral
information each pixel for all pixels. The processor may be
configured to determine a Wiener matrix based on the one or more
reference narrow-band Raman images and the one or more reference
images. The one or more reference images may be generated based on
one or more reference samples, each reference sample including one
or more basic biochemical components. The processor may be
configured to generate the one or more reconstructed Raman images
based on the one or more narrow-band Raman images and the Wiener
matrix. The processor may be configured to remove fluorescence
background from the or more reconstructed Raman images. The one or
more narrow-band Raman images may have a spectral resolution lower
than the one or more reconstructed Raman images. The one or more
filters may be generated from one or more principal components
based on Raman spectra of the reference samples.
[0212] The device may have an interface portion. The lens system
1608 may be configured to direct the light emitted from the optical
source 1604 to the interface portion. The lens system 1608 may be
configured to direct the light from the interface portion to the
detector 1606. The further light may have a frequency shift from
the light. In other words, the frequencies of the light and the
further light may be different.
[0213] The optical source 1604 may be a laser source.
[0214] Methods described herein may further contain analogous
features of any structure, device or array described herein.
Correspondingly, structures, devices or arrays described herein may
further contain analogous features of any method described
herein.
[0215] While the invention has been particularly shown and
described with reference to specific embodiments, it should be
understood by those skilled in the art that various changes in form
and detail may be made therein without departing from the spirit
and scope of the invention as defined by the appended claims. The
scope of the invention is thus indicated by the appended claims and
all changes which come within the meaning and range of equivalency
of the claims are therefore intended to be embraced.
* * * * *