U.S. patent application number 11/430721 was filed with the patent office on 2006-09-14 for system for creating microscopic digital montage images.
Invention is credited to Jeffrey A. Beckstead, Patricia A. Feineigle, John R. II Gilbertson, Christopher R. Hauser, Frank A. JR. Palmieri, Arthur W. Wetzel.
Application Number | 20060204072 11/430721 |
Document ID | / |
Family ID | 25442106 |
Filed Date | 2006-09-14 |
United States Patent
Application |
20060204072 |
Kind Code |
A1 |
Wetzel; Arthur W. ; et
al. |
September 14, 2006 |
System for creating microscopic digital montage images
Abstract
An imaging apparatus. The imaging apparatus may find an area in
which a specimen is present, then focus on the specimen and capture
images of the specimen during continuous stage motion.
Inventors: |
Wetzel; Arthur W.;
(Murrysville, PA) ; Gilbertson; John R. II;
(Pittsburgh, PA) ; Beckstead; Jeffrey A.;
(Valencia, PA) ; Feineigle; Patricia A.;
(Pittsburgh, PA) ; Hauser; Christopher R.;
(Pittsburgh, PA) ; Palmieri; Frank A. JR.;
(Gibsonia, PA) |
Correspondence
Address: |
THE LAW OFFICE OF RICHARD W. JAMES
25 CHURCHILL ROAD
CHURCHILL
PA
15235
US
|
Family ID: |
25442106 |
Appl. No.: |
11/430721 |
Filed: |
May 9, 2006 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
09919452 |
Jul 31, 2001 |
|
|
|
11430721 |
May 9, 2006 |
|
|
|
09757703 |
Jan 11, 2001 |
6798571 |
|
|
09919452 |
Jul 31, 2001 |
|
|
|
09758037 |
Jan 11, 2001 |
6993169 |
|
|
09919452 |
Jul 31, 2001 |
|
|
|
09788666 |
Feb 21, 2001 |
6816606 |
|
|
09919452 |
Jul 31, 2001 |
|
|
|
Current U.S.
Class: |
382/133 |
Current CPC
Class: |
G02B 21/241 20130101;
G02B 21/06 20130101; G02B 21/367 20130101; G01B 7/003 20130101;
G06K 9/3233 20130101; H04N 7/188 20130101; G02B 21/0016 20130101;
G06K 9/00134 20130101; G06T 7/70 20170101 |
Class at
Publication: |
382/133 |
International
Class: |
G06K 9/00 20060101
G06K009/00 |
Claims
1. A focusing method, comprising: focusing a microscope optic on a
plurality of points of a specimen to determine a focal length for
each point; fitting a surface to the points at the focal lengths;
and capturing a plurality of microscopic images of portions of the
specimen while focusing on each microscopic image based on the
fitted surface.
2. The focusing method of claim 1, wherein the fitted surface is
curved.
3. The focusing method of claim 1, wherein the fitted surface is
planar.
4. The focusing method of claim 1, wherein the fitted surface is
described by a model.
5. The focusing method of claim 1, wherein the fitted surface is
described by an equation.
6. The focusing method of claim 1, wherein the fitted surface is a
focal surface.
7. The focusing method of claim 1, wherein the fitted surface
includes at least comprises an area equivalent to an area of the
specimen.
8. The focusing method of claim 1, wherein the captured microscopic
images are adjacent.
9. The focusing method of claim 1, further comprising capturing a
first image of the specimen; and selecting the plurality of points
on which to focus the microscopic optic from the first image.
10. The focusing method of claim 9, wherein the first image is a
macroscopic image.
11. The focusing method of claim 9, wherein the first image is
captured at a lower magnification than that of any of the
microscopic images.
12. The focusing method of claim 9, wherein the plurality of points
are selected based on a relative intensity of the image at each of
the points.
13. The focusing method of claim 1, wherein the plurality of points
are selected based on their distribution across the specimen.
14. The focusing method of claim 1, wherein the specimen is divided
into regions and the plurality of points are each selected from a
different one of the regions.
15. The focusing method of claim 1, further comprising calculating
an error function to determine a fit accuracy of the fitted
surface.
16. The focusing method of claim 1, further comprising: selecting
an additional point; determining another focal length for the
additional point; and fitting another surface to the plurality of
points at the focal lengths and the additional point at the other
focal length.
17. The focusing method of claim 1, wherein the specimen is moved
continuously while a plurality of adjacent images of the specimen
are captured; and wherein the microscope optic is focused based on
the focal length of the fitted surface for each of the adjacent
images to be captured while the specimen is moved.
18. An imaging apparatus, comprising: a motorized stage; a camera
focused relative to the motorized stage; and a processor coupled to
the camera, wherein the processor contains instructions which, when
executed by the processor, cause the processor to: capture a low
resolution image that is incident on the camera, wherein the low
resolution image includes a plurality of pixels, the pixels having
a characteristic; establish the characteristic for each pixel;
determine which of the pixels contain a target image based on the
characteristic of the pixels and establish a target area that
includes those pixels; transpose the position of the target area
into a plurality of stage coordinates; and capture a high
resolution image that is incident on the camera at each of the
stage coordinates.
19. The imaging apparatus of claim 18, wherein the characteristic
includes pixel intensity.
20. The imaging apparatus of claim 18, wherein the characteristic
includes pixel color.
21. The imaging apparatus of claim 18, wherein the processor
determines which pixels contain the target image based on a
relative intensity of the pixels and further: determines a mean
intensity of the pixels; compares the intensity of each of the
pixels to the mean intensity; and divides the pixels into a group
of non-target image pixels having high intensities and a group of
target image pixels having intensities lower than the high
intensities.
22. The imaging apparatus of claim 18, wherein the processor
determines which of the pixels contain the target image based on a
relative intensity of the pixels and further: determines, for the
pixels, an intensity standard deviation that provides an amount of
variation in pixel intensity; compares the intensity of each pixel
to the intensity standard deviation; and divides the pixels into a
group of non-target image pixels having low standard deviations and
a group of target image border pixels having standard deviations
that are greater than any of the low standard deviations.
23. The imaging apparatus of claim 18, further comprising a pulsed
light directed toward the motorized stage.
24. The imaging apparatus of claim 18, further comprising a stage
position sensor adjacent the motorized stage.
25. An imaging apparatus, comprising: a motorized stage; a camera
having a lens directed toward the motorized stage; and a processor
coupled to the camera, wherein the processor contains instructions
which, when executed by the processor, cause the processor to:
select at least three points of a sample adjacent the motorized
stage; determine a stage position for each of the selected points;
focus the camera on each of the selected points; determine an
object distance from the camera lens to the sample at each of the
selected points; and develop a focus surface based on the stage
position and the object distance for the selected points.
26. The imaging apparatus of claim 25, wherein the selecting at
least three points of a sample adjacent the motorized stage
includes selecting points dependent on a characteristic of an image
of the sample at those points.
27. The imaging apparatus of claim 25, wherein the selecting points
dependent on a characteristic of the image of the sample at those
points includes selecting the darkest regions.
28. The imaging apparatus of claim 25, wherein the selecting points
dependent on a characteristic of the image of those regions
includes selecting the lightest regions.
29. The imaging apparatus of claim 25, wherein the selecting points
dependent on a characteristic of the image of those regions
includes selecting points having a high contrast relative to the
regions.
30. The imaging apparatus of claim 25, wherein when the processor
selects at least three points of a sample adjacent the motorized
stage, the processor further: determines a distribution of the at
least three selected points within the sample; determines whether
at least one of the selected points lies within each of at least
two predetermined areas; and selects additional points until at
least one of the additional points lies within each predetermined
area.
31. An imaging apparatus, comprising: a motorized stage; a camera
focused relative to the motorized stage; a stage position sensor
adjacent the motorized stage; and a pulsed light directed toward
the motorized stage and coupled to the stage position sensor such
that the pulsed light illuminates in response to the stage position
sensor.
32. The imaging apparatus of claim 31, further comprising a
processor coupled to the camera, the pulsed light, and the stage
position sensor, wherein the processor contains instructions which,
when executed, cause the processor to: initiate motion of the
motorized stage; energize the pulsed light when the stage position
sensor indicates the motorized stage is in a predetermined
position; and capture an image by way of the camera while the
pulsed light is energized.
33. The imaging apparatus of claim 32, wherein the motorized stage
moves continuously while the images are captured.
Description
CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] The present application is a continuation of U.S. patent
application Ser. No. 09/919,452, filed Jul. 31. 2001, which in turn
is a continuation-in-part of U.S. patent application Ser. No.
09/757,703, filed Jan. 11, 2001, U.S. patent application Ser. No.
09/758,037, filed Jan. 11, 2001, and U.S. patent application Ser.
No. 09/788,666, filed Feb. 21, 2001, all of which are currently
pending and assigned to the assignee of the present invention.
FIELD OF THE INVENTION
[0002] The present invention relates to microscopic digital imaging
of complete tissue sections for medical and research use. In
particular, it describes a method for high throughput montage
imaging of microscope slides using a standard microscope, digital
video cameras, and an illumination system.
BACKGROUND OF THE INVENTION
[0003] Laboratories in many biomedical specialties, such as
anatomic pathology, hematology, and microbiology, examine tissue
under a microscope for the presence and the nature of disease. In
recent years, these laboratories have shown a growing interest in
microscopic digital imaging as an adjunct to direct visual
examination. Digital imaging has a number of advantages including
the ability to document disease, share findings, collaborate (as in
telemedicine), and analyze morphologic findings by computer. Though
numerous studies have shown that digital image quality is
acceptable for most clinical and research use, some aspects of
microscopic digital imaging are limited in application.
[0004] Perhaps the most important limitation to microscopic digital
imaging is a "subsampling" problem encountered in all single frame
images. The sub-sampling problem has two components: a field of
view problem and a resolution-based problem. The field of view
problem limits an investigator looking at a single frame because
what lies outside the view of an image on a slide cannot be
determined. The resolution-based problem occurs when the
investigator looking at an image is limited to viewing a single
resolution of the image. The investigator cannot "zoom in" for a
closer examination or "zoom out" for a bird's eye view when only a
single resolution is available. Significantly, the field of view
and resolution-based problems are inversely related. Thus, as one
increases magnification to improve resolution, one decreases the
field of view. For example, as a general rule, increasing
magnification by a factor of two decreases the field of view by a
factor of four.
[0005] To get around the limitations of single frame imaging,
developers have looked at two general options. The first option
takes the form of "dynamic-robotic" imaging, in which a video
camera on the microscope transmits close to real time images to the
investigator looking at a monitor, while the investigator operates
the microscope by remote control. These systems have been used
successfully in initial telepathology collaborations by allowing a
distant consultant to view the specimen without the delays and
losses associated with sending the physical slide to the consultant
for review, and by allowing the consultant to view the entire
slide, not just a few static images captured by the initial
user.
[0006] However, these systems may not lend themselves to
significant collaborations, documentation or computer based
analysis. To be successful, remote transmission requires lossy
video compression techniques to be used in order to meet the
network bandwidth requirements, or requires significant delays in
the image display if lossless transmission is used. In addition,
lossy compression on the order required for real-time remote
transmission, severely limits computer-based analysis, as well as
human diagnosis, due to the artifacts associated with lossy
compression techniques. Remote operation of a microscope also
requires only a single user to use the instrument at one time,
requiring instrument scheduling and local maintenance of the
instrument and the slides to be viewed.
[0007] The second option being investigated to overcome the
limitations inherit in single frame imaging is a montage (or
"virtual slide") approach. In this method, a robotic microscope
systematically scans the entire slide, taking an image at each
"camera field" corresponding to the field of view of the camera.
Camera field and field of view shall hereinafter be referred to as
the "field." The individual images are then "knitted" together in a
software application to form a very large data set with very
appealing properties. The robotic microscope can span the entire
slide area at a resolution limited only by the power of the optical
system and camera. Software exists to display this data set at any
resolution on a computer screen, allowing the user to zoom in, zoom
out, and pan around the data set as if using a physical microscope.
The data set can be stored for documentation, shared over the
Internet, or analyzed by computer programs.
[0008] The "virtual slide" option has some limitations, however.
One of the limitations is file size. For an average tissue section,
the data generated at 0.33 .mu.m/pixel can be between two and five
gigabytes uncompressed. In an extreme case, the data generated from
one slide can be up to thirty-six gigabytes.
[0009] A much more difficult limitation with the prior systems is
an image capture time problem. Given an optical primary
magnification of twenty and a two-third inch coupled device or
"CCD", the system field of view is approximately (8.8 mm times 6.6
mm)/20=0.44 times 0.33 mm. A standard tissue section of
approximately 2.25 square centimeters, therefore, requires
approximately fifteen hundred fields to capture an image of the
entire tissue section.
[0010] Field rate, which is the amount of time it takes to capture
an image of a field and set-up the apparatus capturing the field
for a following image capture, in montage systems is limited by
three factors--camera frame rate (the number of camera images
acquired per second), image processing speed (including any time
required to read the camera data, perform any processing on the
camera data prior to storage, and to store the final data), and
rate of slide motion, which is the time required for the slide to
be mechanically repositioned for the next image acquisition. Given
today's technology, the rate of slide motion is a significant
limiting factor largely because the existing imaging systems
require the slide to come to a stop at the center of each field to
capture a blur free image of the field.
[0011] For example, traditional bright field microscopic
illumination systems were designed to support direct visual
examination of a specimen on the field and therefore depend on a
continuous light source for illumination. Continuous light,
however, is a significant limitation for digital imaging in that
the slide, which must move to capture an entire image, but must be
stationary with respect to the camera during CCD integration, thus
moving the slide from the light. Moreover, slide motion during
integration results in a blurred image. Traditional montage
systems, therefore, have had to move the slide (and stage) from
field to field in a precise "move, stop, take image and move again"
pattern. This pattern requires precise, expensive mechanics, and
its speed is inherently limited by the inertia of the stage.
[0012] The three-dimensional characteristic of a typical tissue
sample and the slide places additional limitations on the imaging
system. Like all lenses, microscope optics have a finite depth of
field--the distance within which objects will appear to be focused.
A typical depth of field is about 8 microns for a 10.times.
objective, and in general, as the magnification increases, the
depth of field decreases. While microscope slides are polished
glass, the flatness of the slide can vary on the order of 50
microns or more across the slide. The variations in the tissue
sample thickness and any defects associated with placing the sample
on the slide, such as folds in the tissue, also affect the optimal
position of the slide with respect to the imaging optics. The
magnitude of the optimal position and the limited depth of field of
the microscope optics require the focus to be adjusted as the
system moves from field to field. The time to refocus the system at
each field also contributes to the overall capture time of the
montage image.
[0013] Thus, a system is needed to reduce the image capture time.
The system must also enable efficient and high quality imaging of a
microscope slide via a high-resolution slide scanning process.
SUMMARY OF THE INVENTION
[0014] The present invention relates to a method and system for
creating a high throughput montage image of microscope slides. The
system includes an optical system, components that are used in a
pre-scan processing, and components for auto-focusing by enabling
accurate focus control of optical elements without requiring the
stage to be stopped and refocused at each tile location. The
optical system includes at least one camera, a motorized stage for
moving a slide while an image of the slide is captured, a pulsed
light illumination system that optically stops motion on the
motorized stage while allowing continuous physical movement of the
motorized stage, and a stage position detector that controls firing
of the pulsed light illumination system at predetermined positions
of the motorized stage. The components that are used in the
pre-scan processing include an image-cropping component, a tissue
finding component and a scan control component. The image-cropping
component and tissue finding component identify tissue regions on
the slide in the optical system and determine exact locations of
tissue on the slide. The scan control component uses information
about the locations to generate control parameters for the
motorized stage and the camera. The components for auto-focusing
include a focal point selection component, a focal surface
determination component, and a scan component. The focal point
selection component and the focal surface determination component
use the control parameters to ensure that a high-quality montage
image is captured. The scan component is able to capture a
high-quality montage image by maintaining motion of the motorized
stage and synchronization of the optical system. The scan component
controls the stage position to maintain in-focus imaging during the
scanning process without stopping the stage and refocusing at each
location and fires a pulsed-illumination source at the appropriate
position to guarantee image alignment between sequential camera
images.
[0015] Accordingly, it is a benefit of the invention that it
provides a microscopic imaging system for whole slide montage in
which standard microscope optics, off the shelf cameras, a simple
motorized stage, and a pulse light illumination system can be used
to produce precisely aligned image tiles, and acquire these image
tiles at a speed limited primarily by the camera frame rate.
[0016] The present invention uses a strobe light, triggered by a
direct Ronchi ruler or other stage-positioning device, to produce
precisely aligned image tiles that can be made into a montage image
of tissue sections on a microscope slide. Significantly, due to the
short light pulse emitted by a strobe, clear images can be obtained
without stopping the microscope stage. This significantly increases
the image throughput while decreasing the expense and precision
required in the stage mechanics.
[0017] In one embodiment, a strobe arc is placed at the position of
the lamp bulb in a standard microscope system. The camera shutter
is opened and the strobe is fired in response to the position of
the stage as reported by a direct position sensor. If stray light
is minimized, the camera exposure can be much longer than the
strobe flash, allowing low cost cameras to be utilized.
[0018] It is another benefit of the invention to significantly
increase the image throughput of a tiling image system by allowing,
through the use of the strobe light, continuous motion of the slide
under the microscope. The inventive system thus eliminates the need
to stop the microscope stage to capture an image.
[0019] It is another benefit of the invention to reduce the demands
of camera, stage, and strobe synchronization by controlling the
firing of the strobe light based on direct stage position feedback,
thereby substantially reducing the mechanical precision required of
the stage and camera components.
[0020] It is another benefit of the invention to use a pre-scan
process applied to a macroscopic image of the entire slide, to
guide a high-resolution slide scanning process and ensure
high-quality images of the entire specimen are acquired. The
pre-scan process includes an image cropping component, a tissue
finding component, a scan control component, a focus point
selection component, a focal surface determination component, and a
scan component. The image cropping and tissue finding components
identify interesting regions on the slide to be scanned. The focus
point selection and focal surface determination components ensure
that a high quality image is captured by the scanning process, by
enabling accurate focus control to be maintained.
[0021] It is another benefit of the invention to use a
high-resolution slide scanning process to control the operation of
the motorized stage, camera and strobe illumination. This process
utilizes information gathered by the pre-scan process, namely the
imaging regions and focus parameters, to control the positioning of
the stage to image only the regions of interest and to ensure the
individual images are well aligned and in focus.
[0022] Additional features and advantages of the invention will be
apparent from the description that follows, or may be learned by
practice of the invention. The objectives and advantages of the
invention to be realized and attained by the microscopic image
capture system will be pointed out in the written description and
claims hereof as well as the appended drawings.
BRIEF DESCRIPTION OF THE DRAWINGS
[0023] The accompanying drawings, which are included to provide a
further understanding of the invention and are incorporated in and
constitute a part of this specification, illustrate embodiments of
the invention that together with the description serve to explain
the principles of the invention.
[0024] FIG. 1 is a partial isometric view of an embodiment of an
imaging apparatus of the present invention;
[0025] FIG. 1a is a front view of the apparatus of FIG. 1;
[0026] FIG. 1b is a side view of the apparatus of FIG. 1;
[0027] FIG. 1c is a top view of the apparatus of FIG. 1;
[0028] FIG. 2 is a macroscopic image resulting from operation of an
embodiment of the cropping component;
[0029] FIG. 3 illustrates a result of the tissue finding
component;
[0030] FIG. 4 is an overlay of FIGS. 2 and 3 illustrating the
regions of the slide to be imaged;
[0031] FIG. 5 illustrates a result of the focus point selection
component on a sample image;
[0032] FIG. 6 is a generated three-dimensional data set for the
image of FIG. 5; and
[0033] FIG. 7 is a set of charts graphically depicting steps
implemented in an embodiment of the inventive system.
DESCRIPTION OF THE PREFERRED EMOBODIMENTS
[0034] Reference will now be made in detail to the preferred
embodiments of the present invention, examples of which are
illustrated in the accompanying drawings. The following paragraphs
describe the functionality of the inventive system and method for
focus controlled, high throughput montage imaging of microscope
slides using a standard microscope, camera, and a motorized
mechanical stage.
[0035] FIG. 1 illustrates a first embodiment of the imaging
apparatus of the present invention. FIGS. 1a-1c illustrate front,
side, and top views, respectively, of the imaging apparatus
illustrated in FIG. 1. It should be noted that not all components
that may be included in an imaging apparatus are illustrated in
FIGS. 1-1c. For example, a stage support that attaches the stage to
the imaging apparatus is not illustrated so as not to obstruct the
view of other components. In this embodiment, a slide 112 to be
imaged is placed on a thumbnail imaging position in a slide holder
on a motorized stage 102. A single frame image containing the
entire slide may be taken with a macro camera 106. One skilled in
the art will recognize, however, that a suitable image may
alternately be captured by combining multiple images taken by a
microscopic camera such as microscopic camera 104. Microscopic and
macroscopic images may furthermore be captured by use of two
separate cameras or through the use of a single camera. The
low-resolution image may be analyzed by software components,
described below, to determine the locations of tissue on slide 112.
This information can then be used to generate control parameters
for stage 102 and microscopic camera 104 to ensure that the
scanning process captures high quality images of only the tissue
regions, substantially reducing the time to scan an average
slide.
[0036] One skilled in the art will recognize that, although
capturing the single macroscopic image saves time, it is not
necessary for the operation of the invention. However, where a
macroscopic image is to be used, multiple macroscopic images may be
required to generate control parameters to the accuracy required
based on the ratio of the macroscopic to microscopic magnifications
and the camera specifications of each camera, if separate cameras
are utilized.
[0037] Once the regions to be microscopically imaged are
identified, the slide is scanned under the microscope optics. To
facilitate rapid imaging of the slide and to avoid the stop image
reposition delays associated with traditional imaging systems, a
high-speed strobe light is used to optically stop the motion of the
stage, and thus the slide specimen, while allowing continuous stage
motion. It should be apparent to one skilled in the art, that any
pulsed illumination system may be used in place of the high-speed
strobe light. To eliminate overlap or missed tissue between the
microscope images, precise alignment of the stage and camera, along
with accurate stage positioning, and camera and strobe
synchronization, are required. To reduce the camera precision
requirements, a direct stage position sensor is used to control the
firing of the strobe, and thus the camera exposure.
[0038] In this fashion, the camera can be operated with a long
exposure window in comparison to the very short strobe flash,
allowing lower cost components, specifically the stage and camera,
to be utilized.
[0039] Specifically in an embodiment illustrated in FIGS. 1-1c, a
pre-scan processing of the low-resolution or thumbnail image
includes an image cropping component, a tissue finding component, a
scan control component, a focus point selection component, and a
focal surface determination component. Those pre-scanning
components may then be followed by a scan component. The image
cropping component and tissue finding component identify tissue
regions on the slide to be scanned. The scan control component
generates the necessary control parameters to scan only the regions
of interest under the microscope optics. The focus point selection
component and a focal surface determination component ensure that
the scan component captures a high-quality montage image, by
enabling accurate focus control of the optical elements without
requiring the stage to be stopped and refocused at each tile
location, substantially reducing the acquisition time.
[0040] One step in processing the thumbnail image consists of
flat-field correcting the macroscopic thumbnail image using a
similar image that may have been obtained from the same camera and
a blank slide. This removes any spatial light anomalies from the
thumbnail image, which may reduce the efficiency of the tissue
finding component. Depending upon the format, or size, of the
camera and the aspect ratio of the slide, a portion of the image
may contain non-slide objects such as a slide carrier (not shown in
the figures). To remove these features, the thumbnail image may be
cropped to remove non-slide objects, thus retaining only the slide
information.
[0041] The image cropping may be accomplished via a two-pass
process. The first pass in such a process determines an approximate
location of the slide boundary, and the second pass fine-tunes this
estimate. The search for the boundary is conducted over upper and
lower intervals corresponding to the regions expected to contain
the upper and lower slide edges, respectively. The slide or region
of interest may be assumed to be positioned near the center,
vertically, in the thumbnail image. The portion of the image
falling outside of the identified slide boundary is removed. It
should be noted that the cropping component and each of the other
components described herein may operate on either a grayscale or
color image. The image may also be cropped at the estimated edge
locations, and then uniformly reduced in size to produce a small
thumbnail image of the slide for rapid, visual slide
identification.
[0042] Because the slide may not be oriented perfectly horizontally
in the original thumbnail image, the identified slide edges are
likely to lie at an angle. Thus, even after cropping, there may be
remnants of the slide edges or cover slip in the cropped image.
Therefore, the image cropping component attempts to identify pixel
blocks that likely contain these remaining edges and flags these
blocks as edges that will not be considered for high resolution
imaging by the tissue finding component.
[0043] The resulting cropped image generated by the image cropping
component may serve as an input to the tissue finding component.
This component locates regions in the thumbnail image that contain
tissue of interest to a specialist. In order to minimize the time
and storage space required to accomplish high-resolution slide
imaging, the inventive system may capture only those regions of the
slide that contain tissue. This approach may be facilitated by
identifying regions containing tissue in the thumbnail image.
[0044] The tissue finding component identifies tissue regions via a
sequence of filters that incorporate knowledge of the typical
appearance and location of tissue and non-tissue slide regions.
Initial filtering steps, in one embodiment, convert the image to a
grayscale image and analyze the mean and standard deviation of the
local pixel intensities. Pixel mean intensities may be used to
differentiate tissue-containing regions from blank and other
non-tissue regions, such as those containing the slide label or
other markings. The standard deviation data may represent the
amount of variation in pixel intensity values and thus is a good
indicator of the border between tissue and the blank slide. The
mean and standard deviation data is combined to generate a
threshold value that is used to make an initial classification of
tissue versus non-tissue. Subsequently, morphological filters may
be applied to refine the classification based on the size and
position of neighboring groups of potential tissue pixels.
[0045] The embodiment described above uses the mean and standard
deviation of the local pixels as the basis for detecting regions of
interest. One skilled in the art will recognize, however, that
other image characteristics can also be used to identify the
specimen from non-items of interest such as dust and scratches.
That embodiment may also process a grayscale macroscopic image. It
should be noted, however, that the pixel intensity differentiation
tools described herein can be applied to each of the color
components (traditionally, red, green and blue) of a color image in
addition to being applied to a grayscale image. Additional
processing tools can also be applied between color components to
refine the tissue finding accuracy and to remove features such as
labels and writing that are not critical to the application, or to
select user defined areas of interest to be scanned, such as
regions circled by a purple marker.
[0046] The filters, which comprise the tissue finding component,
process the pixels of the cropped grayscale thumbnail image in
groups that correspond to slide regions, or tiles, that can be
imaged individually during the high-resolution scanning process.
These filters ensure that tiles only partially filled with tissue
are classified as tissue-containing tiles. The final output of the
filter sequence is a tiling matrix, the value of which indicates
which tiles should be imaged. The tiling matrix subsequently guides
the high-resolution scanning process.
[0047] An example of the image cropping and tissue finding
processes are shown in FIGS. 2, 3 and 4. FIG. 2 illustrates the
macroscopic image after flat-field correction and image cropping.
FIG. 3 illustrates the results of the tissue finding component. The
resulting tile matrix shown in FIG. 3 has a one-to-one
correspondence to the field of view of the microscopic camera.
White pixels (binary 1) signify fields to be captured and imaged,
and black pixels represent regions not to image. FIG. 4 illustrates
an overlay of FIGS. 2 and 3 representing the sections of the slide
to be imaged. For the application depicted in FIGS. 2-4 (anatomical
pathology), it may be important to image all suspect regions that
may contain tissue so conservative criteria are used in the tissue
finding component, resulting in cover slip edges and writing etched
into the slide to be identified as to be imaged. The savings in the
acquisition time is represented by the ratio of the white to black
areas of FIG. 3. For this image, only 53% of the slide region is to
be imaged, including the label and cover slip edges, and etched
writing on the slide.
[0048] At the completion of the tissue finding component, the scan
control component interprets the tissue finding tile matrix (FIG.
3) and transposes the positions into actual stage coordinates for
the microscopic imaging. A program running on a host computer
controls the operation by communicating with a stage controller and
microscopic camera 104. Actual scanning can occur in any fashion
such as by rows or columns, or in a step fashion to image
neighboring areas.
[0049] To achieve good focus for an entire slide, the surface that
best represents the focal position of the sample with respect to
the optical elements may be determined and used to automatically
control the focus as the tissue is scanned under the microscope
optics. These steps may furthermore be completed under the focus
point selection component and the focus surface determination
component.
[0050] In one embodiment, the focus point selection component
evaluates the tissue regions of the thumbnail image and selects
several points on which to initially focus the microscope optics.
In this embodiment, that focus point selection is based on pixel
intensity and distribution. Initially, tissue containing pixels or
groups of tissue containing pixels are considered. Those pixels or
pixel groups will be referred to hereinafter as regions. The
regions that will be focused upon, or focus regions, are then
selected based on the relative contrast between the intensity of
those regions within the thumbnail image and the distribution of
those regions within the tissue containing portions of the image.
It will be recognized that other intensity and/or distribution
selection criterion may be utilized to select focus regions. After
the focus regions are identified, a normalized focus surface such
as a plane or curve may be drawn through the focus points and an
equation, model, curve, or other method for describing appropriate
focal lengths along that surface can be obtained. Proper focus for
each region to be captured by the high-resolution scan may then be
calculated by focusing on, for example, the plane or curved
surface.
[0051] Selecting the appropriate focus region based on relative
contrast improves the likelihood that regions that contain tissue
will be selected over a method of picking focus regions based on a
pre-selected grid-pattern. Selecting regions based on their overall
distribution with respect to the tissue coverage area provides
assurance that the resulting plane or surface will be
representative of the entire slide and not limited to a small
portion of the tissue, as could occur if selection were based
solely on the relative contrast of pixels.
[0052] To select focus regions based on contrast, the present
invention may select a number of focus regions having a desired
contrast quality. For example, six to ten of the darkest regions
may be selected from a back-lit image under the assumption that
dark regions contain a large amount of tissue.
[0053] Distribution refers to the overall distribution of regions
with respect to the tissue coverage area. Because a surface defined
by the focus regions is the basis for maintaining the focus across
the entire specimen during scanning, it is beneficial to have focus
regions dispersed across the specimen rather than being grouped in
close proximity to one another in one or more areas of the
specimen. In that way, every point to be scanned will be close to a
focus point. Thus, the scanned optical position at each point, as
defined by the surface, should be very nearly the optimum in-focus
position for that point.
[0054] The focus point selection component may, for example, assure
that at least one focus point is located within each of a number of
pre-selected areas of the tissue. Those areas may, for example, be
separate pieces of tissue on a slide. If a focus point does not
exist on each pre-selected area, the focus point selection
component may select additional regions, for example, having the
desired contrast quality, until at least one focus point is
identified on each pre-selected area. The number of data points
required will depend on the actual three-dimensional structure
defined by the specimen and slide, and the geometrical dimension of
the surface to be fit. Once the surface has been determined, an
error function can be calculated to determine the fit accuracy. For
example, the mean square error of each selected focal region may be
calculated to determine how much error exists between the surface
fit and each focus region. If that error is greater than a
predetermined acceptable error level, at one or more points,
additional data points may be added to the calculation and/or
points that have large errors may be eliminated and the surface may
be recalculated under the assumption that the points having
excessive errors were anomalies. A surface may then be fitted to
those points to define a focal surface to be utilized when scanning
the tissue.
[0055] In alternative embodiments, the focus points are either user
definable through an input file or through a suitable user
interface. In addition, for cases where the specimen locations are
reproducible on the slides, the focus points can be predefined and
repeated for each slide without the use of a macroscopic image or
any preprocessing to find the tissue regions.
[0056] Once selected, each focus position is placed under the
microscope optics in the focal surface determination component, and
an auto-focus program determines the best-fit focus at each
position. This generates a three-dimensional data set corresponding
to the optimal specimen distance at each stage location. These data
points are used as input to a surface fitting routine that
generates the necessary control parameters for the slide scanning
process.
[0057] At the completion of the focus point selection, the tissue
information and the surface parameters are passed to the scan
control component. This component is responsible for the motion of
the stage and synchronization of the microscopic imaging system
during montage image acquisition. To achieve accurate, well-aligned
tiled images, the specimen may be positioned such that each camera
image is aligned within the equivalent single pixel spacing in real
or stage space (camera pixel size divided by the overall optical
magnification). This usually entails a stage step of ox and by
where each step is directly related to the effective pixel size,
which may be expressed in terms of camera pixel size/optical
magnification, and the number of image pixels in the x and y
directions respectively. For example, a 1300.times.1030 pixel, 10
.mu.m square pixel camera operated at 20.times. magnification
results in .delta.x=10 .mu.m*1300/20=650 .mu.m and .delta.y=10
.mu.m*1030/20=515 .mu.m. To maintain focus during the scanning
process, the stage may be positioned at the proper focal position
as determined by the focus surface parameters: z.sub.ij=f(x.sub.i,
y.sub.j), where z.sub.ij is the vertical position of the slide with
respect to the optical components, f(x.sub.i, y.sub.j) is the
function used to represent the best focus surface, and x.sub.i and
y.sub.j are the positions of each camera image in the x and y axes,
respectively. The camera image positions can be expressed as linear
relations between the starting position of the stage (x.sub.0 and
y.sub.o) and the step size associated with each image dimension:
x.sub.i=x.sub.0+i*.delta.x and y.sub.i=y.sub.0+i*.delta.y.
[0058] Image montage scanning is traditionally accomplished by
either scanning by rows or columns. Assuming that tiling is
completed by scanning across a row in the x-direction and stepping
vertically in the y-direction after each row has been scanned, the
stage is simply positioned at the appropriate position given by
z.sub.i,j=f(x.sub.i, y.sub.j). Thereafter, the stage is stopped and
an image is acquired. If imaging is accomplished during continuous
motion of the stage in the x-direction via a line scan camera or
alternative imaging arrangement, the vertical velocity as a
function of focal position x.sub.i and time, can be computed from
the partial derivative of the focal surface: V.sub.z(v.sub.x,
y.sub.j)=..delta.f(x.sub.i, y.sub.j)/.delta.x*v.sub.x, where
V.sub.z is the velocity of the vertical position of the stage and
v.sub.x is the velocity in the x-direction. The velocity of the
vertical position of the stage can be used to control the optical
position and maintain focus as images are acquired continuously
across the row.
[0059] FIG. 5 represents the results of the focus point selection
component. This figure shows the thumbnail or macroscopic image of
the region to be scanned. The light spots 504 overlaid on the
specimen 506 represent the positions selected by the focus point
selection component. These positions are placed under the
microscope and auto-focused on each location. FIG. 6 illustrates
the three-dimensional data set generated by focusing on each of the
focus points of the specimen depicted in FIG. 5. For this slide,
the best fit was planar, z(x,y)=dz/dx x+dz/dy y+z0, where dz/dx
(dz/dy) is the slope or pitch of the plane with respect to the
x-axis (y-axis) and z0 is the vertical offset of the plane with
respect to the z-axis. The best fit parameters for the specimen of
FIG. 5 are also shown in FIG. 6.
[0060] At the completion of the pre-scan processing, the tile
matrix and the stage control parameters are passed to a
scanning-process control program. The scanning process control
program controls the operation by communicating with a stage
controller, a stage position sensor, a camera and the strobe firing
circuitry. In the invention, the computer program controls the
operation of stage 102, camera 104 and strobe 108 illumination. The
actual slide scanning can be automated to image entire slides,
image only a portion of the slide or use a user-interface to allow
the user to select the regions to be imaged. Once a region has been
selected for imaging, the program then controls the operation by
communicating with a stage controller, a stage position sensor,
camera 104 and strobe firing circuitry 108. Preferably, tiling is
performed by moving stepwise along the short axis and with
continuous motion along the long axis. In other words, tiling is
done one row at a time. For this reason, a stage position is
monitored and controlled differently along each stage axis. Along
the short axis of the slide, the stage position is monitored and
controlled, by the program, directly through the stage controller.
Along the long axis, however, the stage position is monitored by a
direct stage position sensor, which can be separate or part of the
overall stage control circuitry.
[0061] In another embodiment, a Ronchi ruler attached to stage 102
is used for the stage position sensor. Those skilled in the art
will recognize that any position sensor may be used in the
invention. This sensor can be external to the stage controller or
the positional information can be acquired directly from the stage
controller with or without feedback.
[0062] For reference, a Ronchi ruler is a pattern of alternating
light and dark bands, equally spaced along a substrate, typically
either glass or plastic. A position sensor based on the Ronchi
ruler utilizes a light sensor that is mechanically isolated from
the ruler. As the ruler passes under the sensor, a series of
electronic pulses is generated corresponding to the alternating
light and dark bands of the ruler. These pulses can be used to
monitor the position and direction of stage 102.
[0063] Based on the magnification of the optics and the camera
utilized, strobe 108 is fired whenever the position sensor
determines stage 102 has moved into the field of view of the next
tile to be captured by the camera 104. The system continues to
capture image tiles with precise alignment, until the images of all
desired files have been captured or the controlling program tells
the system to stop. At the end of the capture process, the slide is
removed and another slide can be inserted. With current technology,
the rate-limiting step for image capture utilizing the present
invention is the data transfer period in the camera.
[0064] FIG. 7 illustrates the signals of camera 104, stage 102,
optical position detector, and strobe 108. Note that in FIG. 7,
graphs 202 and 204 the signals from the optical position detector
represent motion of stage 102, so their timing will vary depending
on the speed of the stage movement. Where the system is triggered
by the location of stage 102 as reported by the optical position
sensor, precise movement of the stage movement is not necessary,
allowing for the use of low cost stages 102.
[0065] The system can be run in a variety of modes, depending on
how the camera is controlled. In one embodiment, the stage
location, as sensed by a position sensor, fires both the camera 104
and the strobe 108 as indicated by the two traces at 204. In an
alternate embodiment, camera 104 is free running and only strobe
108 is fired by stage position as indicated by a single trace at
204. This mode does not depend on uniform motion of stage 102 over
the area imaged, because the strobe pulse is much shorter than the
integration time of the camera 104, wherein the integration time is
a time during which the camera is receiving an image. As long as
the correct stage position is reached anytime within the
integration time of camera 104, an excellent, well aligned image
results.
[0066] Firing strobe 108 based on direct position information
differs from the more traditional application of strobe
photography. In traditional strobe photography, a strobe and camera
are synchronized in time, and positional information regarding the
objects being imaged can be inferred from the relative position
within the image. When the present invention is operated in a mode
wherein the position feedback controls both camera 104 and strobe
108, and camera 104 is not free running, each camera frame
corresponds to an equally spaced positional change, independent of
the stage velocity (speed and time variations in the speed). In the
case that camera 104 is free running, the stage speed has to be
matched to the camera frame rate only to the accuracy such that the
strobe pulse does not fall outside the exposure window. The
relative time within the exposure window is irrelevant.
[0067] As will be understood by one skilled in the art, while the
present invention describes a microscopic optical arrangement, the
invention can also be applied to other optical imaging, inspection
and illumination systems that are used for building up an image by
matching the stage speed with the camera speed.
[0068] The foregoing description has been directed to specific
embodiments of this invention. It will be apparent, however, that
other variations and modifications may be made to the described
embodiments, with the attainment of some or all of their
advantages. Therefore, it is the object of the appended claims to
cover all such variations and modifications as come within the true
spirit and scope of the invention.
* * * * *