U.S. patent application number 16/446174 was filed with the patent office on 2020-01-02 for video decoding method, video decoder, video encoding method and video encoder.
This patent application is currently assigned to INDUSTRIAL TECHNOLOGY RESEARCH INSTITUTE. The applicant listed for this patent is INDUSTRIAL TECHNOLOGY RESEARCH INSTITUTE. Invention is credited to Ching-Chieh LIN, Chun-Lung LIN, Po-Han LIN, Sheng-Po WANG, Chang-Hao YAU.
Application Number | 20200007872 16/446174 |
Document ID | / |
Family ID | 67105955 |
Filed Date | 2020-01-02 |
![](/patent/app/20200007872/US20200007872A1-20200102-D00000.png)
![](/patent/app/20200007872/US20200007872A1-20200102-D00001.png)
![](/patent/app/20200007872/US20200007872A1-20200102-D00002.png)
![](/patent/app/20200007872/US20200007872A1-20200102-D00003.png)
![](/patent/app/20200007872/US20200007872A1-20200102-D00004.png)
![](/patent/app/20200007872/US20200007872A1-20200102-D00005.png)
![](/patent/app/20200007872/US20200007872A1-20200102-D00006.png)
![](/patent/app/20200007872/US20200007872A1-20200102-D00007.png)
![](/patent/app/20200007872/US20200007872A1-20200102-D00008.png)
![](/patent/app/20200007872/US20200007872A1-20200102-D00009.png)
![](/patent/app/20200007872/US20200007872A1-20200102-D00010.png)
View All Diagrams
United States Patent
Application |
20200007872 |
Kind Code |
A1 |
WANG; Sheng-Po ; et
al. |
January 2, 2020 |
VIDEO DECODING METHOD, VIDEO DECODER, VIDEO ENCODING METHOD AND
VIDEO ENCODER
Abstract
A video decoding method includes: receiving a coding value; and
performing the following steps according to an index value of the
coding value: collecting a plurality of reference samples, grouping
the plurality of reference samples to generate at least one group,
establishing a model of the at least one group, obtaining a target
pixel from a target block, selecting a target group from the at
least one group, and introducing a luminance value of the target
pixel into a model of the target group to predict a chromaticity
value of the target pixel.
Inventors: |
WANG; Sheng-Po; (Taoyuan
City, TW) ; LIN; Chun-Lung; (Taipei City, TW)
; LIN; Ching-Chieh; (Taipei City, TW) ; YAU;
Chang-Hao; (New Taipei City, TW) ; LIN; Po-Han;
(Taipei City, TW) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
INDUSTRIAL TECHNOLOGY RESEARCH INSTITUTE |
Hsinchu |
|
TW |
|
|
Assignee: |
INDUSTRIAL TECHNOLOGY RESEARCH
INSTITUTE
Hsinchu
TW
|
Family ID: |
67105955 |
Appl. No.: |
16/446174 |
Filed: |
June 19, 2019 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
62691729 |
Jun 29, 2018 |
|
|
|
62727595 |
Sep 6, 2018 |
|
|
|
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
H04N 19/176 20141101;
H04N 19/11 20141101; H04N 19/70 20141101; H04N 19/593 20141101;
H04N 19/105 20141101; H04N 19/132 20141101; H04N 19/159 20141101;
H04N 19/117 20141101 |
International
Class: |
H04N 19/176 20060101
H04N019/176; H04N 19/105 20060101 H04N019/105; H04N 19/593 20060101
H04N019/593 |
Claims
1. A video decoding method, comprising: receiving a coding value,
and performing steps below according to an index value of the
coding value: collecting a plurality of reference samples; grouping
the reference samples to generate at least one group; establishing
a model of the at least one group; obtaining a target pixel from a
target block; selecting a target group from the at least one group
according to the target pixel; and introducing a luminance value of
the target pixel into a model of the target group to predict a
chromaticity value of the target pixel.
2. The video decoding method according to claim 1, wherein a
quantity of the groups generated is one.
3. The video decoding method according to claim 1, wherein a
quantity of the groups generated is any number more than two.
4. The video decoding method according to claim 1, wherein the step
of grouping the reference samples comprises: calculating an average
luminance value of the reference samples; and assigning, according
to whether respective luminance values of the reference samples are
greater than or smaller than the average luminance value, the
reference samples into the at least one generated group.
5. The video decoding method according to claim 1, wherein the step
of grouping the reference samples comprises: establishing a first
group according to a first reference sample of the reference
samples, and assigning the first reference sample into the first
group; determining, according to a sample characteristic value of a
second reference sample of the reference samples, whether to add
the second reference sample into the first group; if it is
determined to add the second reference sample into the first group,
updating a group characteristic value of the first group; and if it
is determined not to add the second reference sample into the first
group, establishing a second group and calculating a group
characteristic value of the second group.
6. The video decoding method according to claim 5, wherein the
group characteristic value comprises any combination of: a
position, a representative luminance value, a representative
chromaticity component value, a maximum luminance, a minimum
luminance, a maximum chromaticity, a minimum chromaticity and a
reference sample quantity; and the sample characteristic value
comprises any combination of: a luminance, at least one
chromaticity component and a position.
7. The video decoding method according to claim 1, wherein the step
of grouping the reference samples comprises: defining a quantity of
the at least one group; establishing the at least one group of a
fixed quantity, and calculating individual group characteristic
values of the at least one group according to individual sample
characteristic values of a plurality of reference samples of the at
least one group; and assigning the reference samples into the at
least one group according to the individual sample characteristic
values.
8. The video decoding method according to claim 7, wherein the
group characteristic value comprises any combination of: a
position, a representative luminance value, a representative
chromaticity component value, a maximum luminance, a minimum
luminance, a maximum chromaticity, a minimum chromaticity and a
reference sample quantity; and the sample characteristic value
comprises any combination of: a luminance, at least one
chromaticity component and a position.
9. The video decoding method according to claim 1, wherein the step
of establishing the model of the at least one group comprises:
establishing the model by applying a linear regression
algorithm.
10. The video decoding method according to claim 1, wherein the
step of establishing the model of the at least one group comprises:
establishing the model by applying a straight line equation.
11. The video decoding method according to claim 1, wherein the
step of establishing the model of the at least one group comprises:
establishing the model by applying an averaging algorithm.
12. The video decoding method according to claim 1, wherein the
step of selecting the target group from the at least one group
comprises: identifying from the at least one group the target group
nearest to the target pixel according to the luminance value of the
target pixel.
13. The video decoding method according to claim 1, wherein, if a
luminance value of a previously processed target pixel is near the
luminance value of the target pixel, the chromaticity value of the
target pixel is predicted by using a predicted chromaticity value
of the previously processed target pixel.
14. The video decoding method according to claim 1, wherein: a
spatial position of a characteristic value discontinuity is
identified to segment a reference block and the target block,
wherein the reference block is segmented into at least one
reference sub-block, and a group is established for the at least
one reference sub-block; and at least one model is established by
using the at least one reference sub-block.
15. The video decoding method according to claim 1, wherein the
step of predicting the chromaticity value of the target pixel
comprises: segmenting the target block into at least one target
sub-block according to a spatial position of a characteristic value
discontinuity; selecting the target group according to the target
sub-block to which the target pixel belongs; and predicting the
chromaticity value of the target pixel by using the model of the
target group.
16. The video decoding method according to claim 1, wherein the
reference sample is collected from a left adjacent block and an
upper adjacent block of the target block.
17. A video encoding method, comprising: collecting a plurality of
reference samples; grouping the reference samples to generate at
least one group; establishing a model of the at least one group;
obtaining a target pixel from a target block; selecting a target
group from the at least one group according to the target pixel;
introducing a luminance value of the target pixel into a model of
the target group to predict a chromaticity value of the target
pixel; and generating a coding value, the coding value having an
index value.
18. The video encoding method according to claim 17, wherein a
quantity of the groups generated is one.
19. The video encoding method according to claim 17, wherein a
quantity of the groups generated is any number more than two.
20. The video encoding method according to claim 17, wherein the
step of grouping the reference samples comprises: calculating an
average luminance value of the reference samples; and assigning,
according to whether respective luminance values of the reference
samples are greater than or smaller than the average luminance
value, the reference samples into the at least one generated
group.
21. The video encoding method according to claim 17, wherein the
step of grouping the reference samples comprises: establishing a
first group according to a first reference sample of the reference
samples, and assigning the first reference sample into the first
group; determining, according to a sample characteristic value of a
second reference sample of the reference samples, whether to add
the second reference sample into the first group; if it is
determined to add the second reference sample into the first group,
updating a group characteristic value of the first group; and if it
is determined not to add the second reference sample into the first
group, establishing a second group and calculating a group
characteristic value of the second group.
22. The video encoding method according to claim 21, wherein the
group characteristic value comprises any combination of: a
position, a representative luminance value, a representative
chromaticity component value, a maximum luminance, a minimum
luminance, a maximum chromaticity, a minimum chromaticity and a
reference sample quantity; and the sample characteristic value
comprises any combination of: a luminance, at least one
chromaticity component and a position.
23. The video encoding method according to claim 17, wherein the
step of grouping the reference samples comprises: defining a
quantity of the at least one group; establishing the at least one
group of a fixed quantity, and calculating individual group
characteristic values of the at least one group according to
individual sample characteristic values of a plurality of reference
samples of the at least one group; and assigning the reference
samples into the at least one group according to the individual
sample characteristic values.
24. The video encoding method according to claim 23, wherein the
group characteristic value comprises any combination of: a
position, a representative luminance value, a representative
chromaticity component value, a maximum luminance, a minimum
luminance, a maximum chromaticity, a minimum chromaticity and a
reference sample quantity; and the sample characteristic value
comprises any combination of: a luminance, at least one
chromaticity component and a position.
25. The video encoding method according to claim 17, wherein the
step of establishing the model of the at least one group comprises:
establishing the model by applying a linear regression
algorithm.
26. The video encoding method according to claim 17, wherein the
step of establishing the model of the at least one group comprises:
establishing the model by applying a straight line equation.
27. The video encoding method according to claim 17, wherein the
step of establishing the model of the at least one group comprises:
establishing the model by applying an averaging algorithm.
28. The video encoding method according to claim 17, wherein the
step of selecting the target group from the at least one group
comprises: identifying from the at least one group the target group
nearest to the target pixel according to the luminance value of the
target pixel.
29. The video encoding method according to claim 17, wherein, if a
luminance value of a previously processed target pixel is near the
luminance value of the target pixel, the chromaticity value of the
target pixel is predicted by using a predicted chromaticity value
of the previously processed target pixel.
30. The video encoding method according to claim 17, wherein: a
spatial position of a characteristic value discontinuity is
identified to segment a reference block and the target block,
wherein the reference block is segmented into at least one
reference sub-block, and a group is established for the at least
one reference sub-block; and at least one model is established by
using the at least one reference sub-block.
31. The video encoding method according to claim 17, wherein the
step of predicting the chromaticity value of the target pixel
comprises: segmenting the target block into at least one target
sub-block according to a spatial position of a characteristic value
discontinuity; selecting the target group according to the target
sub-block to which the target pixel belongs; and predicting the
chromaticity value of the target pixel by using the model of the
target group.
32. The video encoding method according to claim 17, wherein the
reference sample is collected from a left adjacent block and an
upper adjacent block of the target block.
33. A video decoder, comprising: a processor, for controlling the
video decoder; a memory, for storing a plurality of reference
samples and a target block; a decoding module; and an index
receiving module, receiving a coding value; wherein, the processor,
the memory, the decoding module and the index receiving module are
coupled to one another; and the decoding module performs operations
below according to an index value of the coding value: collecting a
plurality of reference samples; grouping the reference samples to
generate at least one group; establishing a model of the at least
one group; obtaining a target pixel from a target block; selecting
a target group from the at least one group according to the target
pixel; and introducing a luminance value of the target pixel into a
model of the target group to predict a chromaticity value of the
target pixel.
34. A video encoder, comprising: a processor, for controlling the
video encoder; a memory, for storing a plurality of reference
samples and a target block; an encoding module; and an index
selecting module; wherein, the processor, the memory, the encoding
module and the index selecting module are coupled to one another;
and the encoding module performs operations below: collecting a
plurality of reference samples; grouping the reference samples to
generate at least one group; establishing a model of the at least
one group; obtaining a target pixel from a target block; selecting
a target group from the at least one group according to the target
pixel; introducing a luminance value of the target pixel into a
model of the target group to predict a chromaticity value of the
target pixel; and generating a coding value; and the index
selecting module generates an index value from the coding value
generated by the encoding module.
Description
[0001] This application claims the benefit of U.S. provisional
application Ser. No. 62/691,729, filed Jun. 29, 2018 and U.S.
provisional application Ser. No. 62/727,595, filed Sep. 6, 2018,
the subject matters of which are incorporated herein by
references.
TECHNICAL FIELD
[0002] The disclosure relates to a video decoding method, a video
decoder, a video encoding method and a video encoder.
BACKGROUND
[0003] To enhance coding efficiency of video data, in international
video coding standards, such as H.264/Advanced Video Coding (AVC),
intra prediction is introduced to remove spatial information
redundancy of a current coding image block and neighboring coded
image blocks.
[0004] When video data in a coding format of YCbCr is coded, if the
amount of coding data is further reduced, encoding/decoding
efficiency can be enhanced.
SUMMARY
[0005] According to an exemplary embodiment of the disclosure, a
video decoding method is provided, the method including: receiving
a coding value; and performing the following steps according to an
index value of the coding value: collecting a plurality of
reference samples, grouping the reference samples to generate at
least one group, establishing a model of the at least one group,
obtaining a target pixel from a target block, selecting a target
group from the at least one group according to the target pixel,
and introducing a luminance value of the target pixel into a model
of the target group to predict a chromaticity value of the target
pixel.
[0006] According to an exemplary embodiment of the disclosure, a
video encoding method is provided, the method including: collecting
a plurality of reference samples, grouping the reference samples to
generate at least one group, establishing a model of the at least
one group, obtaining a target pixel from a target block, selecting
a target group from the at least one group according to the target
pixel, introducing a luminance value of the target pixel into a
model of the target group to predict a chromaticity value of the
target pixel, and generating a coding value, the coding value
including an index value.
[0007] According to an exemplary embodiment of the disclosure, a
video decoder is provided, the decoder including: a processor, for
controlling the video decoder; a memory, for storing a plurality of
reference samples and a target block; a decoding module; and an
index receiving module, receiving a coding value. The processor,
the memory, the decoding module and the index receiving module are
coupled to one another. The decoding module performs the following
steps according to an index value of the coding value: collecting a
plurality of reference samples, grouping the reference samples to
generate at least one group, establishing a model of the at least
one group, obtaining a target pixel from a target block, selecting
a target group from the at least one group according to the target
pixel, and introducing a luminance value of the target pixel into a
model of the target group to predict a chromaticity value of the
target pixel.
[0008] According to an exemplary embodiment of the disclosure, a
video encoder is provided, the encoder including: a processor, for
controlling the video encoder; a memory, for storing a plurality of
reference samples and a target block; an encoding module; and an
index selecting module. The processor, the memory, the encoding
module and the index selecting module are coupled to one another.
The encoding module performs the following operations: collecting
the reference samples, grouping the reference samples to generate
at least one group, establishing a model of the at least one group,
obtaining a target pixel from the target block, introducing a
luminance value of the target pixel into a model of the target
group to predict a chromaticity value of the target pixel, and
generating a coding value. The index selecting module generates an
index value from the coding value generated by the encoding
module.
BRIEF DESCRIPTION OF THE DRAWINGS
[0009] FIG. 1 is a flowchart of a video decoding method according
to an exemplary embodiment of the disclosure;
[0010] FIG. 2A to FIG. 2C are reference blocks according to an
exemplary embodiment of the disclosure;
[0011] FIG. 3 is a schematic diagram of grouping according to an
exemplary embodiment of the disclosure;
[0012] FIG. 4 is a schematic diagram of grouping according to an
exemplary embodiment of the disclosure;
[0013] FIG. 5 is a schematic diagram of grouping according to an
exemplary embodiment of the disclosure;
[0014] FIG. 6 is a schematic diagram of segmentation according to
an exemplary embodiment of the disclosure;
[0015] FIG. 7 is a schematic diagram of another type of
segmentation according to an exemplary embodiment of the
disclosure;
[0016] FIG. 8 is a schematic diagram of establishing a linear model
according to an exemplary embodiment of the disclosure;
[0017] FIG. 9A to FIG. 9C are schematic diagrams of establishing a
linear model according to an exemplary embodiment of the
disclosure;
[0018] FIG. 10 is a schematic diagram of establishing a linear
model according to an exemplary embodiment of the disclosure;
[0019] FIG. 11 is a schematic diagram of prediction of a
chromaticity value according to an exemplary embodiment of the
disclosure;
[0020] FIG. 12 is a schematic diagram of prediction of a
chromaticity value according to an exemplary embodiment of the
disclosure, wherein the target pixel is regarded as an
out-tier;
[0021] FIG. 13 is a flowchart of a video encoding method according
to an exemplary embodiment of the disclosure;
[0022] FIG. 14 is a function block diagram of a video decoder
according to an exemplary embodiment of the disclosure; and
[0023] FIG. 15 is a function block diagram of a video encoder
according to an exemplary embodiment of the disclosure.
[0024] In the following detailed description, for purposes of
explanation, numerous specific details are set forth in order to
provide a thorough understanding of the disclosed embodiments. It
will be apparent, however, that one or more embodiments may be
practiced without these specific details. In other instances,
well-known structures and devices are schematically shown in order
to simplify the drawing.
DETAILED DESCRIPTION
[0025] Technical terms of the application are based on general
definition in the technical field of the application. If the
application describes or explains one or some terms, definition of
the terms are based on the description or explanation of the
application. Each of the disclosed embodiments has one or more
technical features. In possible implementation, one skilled person
in the art would selectively implement part or all technical
features of any embodiment of the application or selectively
combine part or all technical features of the embodiments of the
application based on the disclosure of the application and his/her
own need.
[0026] FIG. 1 shows a flowchart of a video decoding method
according to an exemplary embodiment of the disclosure. As shown in
FIG. 1, in step 105, a coding value is received. The following
steps are performed according to an index value of the coding
value. In step 110, a plurality of reference samples are collected.
In step 120, the reference samples are grouped to generate at least
one group, wherein the quantity of groups generated is one or any
number more than two. In step 130, a model of the at least one
group is established (for example but not limited to, a correlation
model). In step 140, a target pixel is obtained (or selected) from
a target block. In step 150, a target group is selected from the at
least one group according to the target pixel. In step 160, a
luminance value of the target pixel is introduced into a model of
the target group to predict a chromaticity value of the target
pixel.
[0027] FIG. 2A to FIG. 2C show reference blocks according to an
exemplary embodiment of the disclosure. As shown in FIG. 2A, each
of reference blocks RB1 and RB2 is taken from one pixel line
adjacent to a target block TB, wherein the reference block RB1 is
obtained from an upper block adjacent to the target block TB and
the reference block RB2 is obtained from a left block adjacent to
the target block RB; that is, the reference samples are collected
from the upper adjacent block and the left adjacent block of the
target block. The target block TB includes 2B*2B pixels (where B is
a positive integer). As shown in FIG. 2B, the reference blocks RB1
and RB2 are taken from two pixel lines adjacent to the target block
TB. As shown in FIG. 2C, the reference blocks RB1 and RB2 are taken
from four pixels lines adjacent to the target block TB.
[0028] The target block TB refers to a block to be reconstructed,
wherein a luminance value of the target block TB is known. In an
exemplary embodiment of the disclosure, a chromaticity value of a
target pixel of the target block TB can be predicted from a
luminance value of the target pixel of the target block TB. After
individual chromaticity values of all of the target pixels in the
target block TB have been predicted, the target block TB is
considered as having been reconstructed. In the description below,
a linear model is taken as an example for illustration; however,
the present application is not limited thereto.
[0029] In step 110, the reference samples are scanned to obtain a
plurality of luminance values and a plurality of chromaticity
values of the reference samples.
[0030] Principles for the grouping in step 120 are described
herein. FIG. 3 shows a schematic diagram of grouping according to
an exemplary embodiment of the disclosure. As shown in FIG. 3, all
reference samples (the dots in FIG. 3 are the reference samples)
are grouped into one single group G1, and a linear model LM1 is
established for the group G1. Then, the luminance value of each
target pixel is introduced into the linear model LM1 to predict the
chromaticity value of each target pixel.
[0031] FIG. 4 shows a schematic diagram of grouping according to an
exemplary embodiment of the disclosure. As shown in FIG. 4, an
average luminance value AL of all of the reference samples is
calculated. All of the reference samples are grouped into two
groups G1 and G2 according to the average luminance value AL,
wherein a reference sample having a luminance value greater than
the average luminance value AL is grouped into the group G1, and a
reference sample having a luminance value smaller than the average
luminance value AL is grouped into the group G2. Linear models LM1
and LM2 are respectively generated for the groups G1 and G2. The
group nearer to the target pixel is identified, so as to predict
the chromaticity value of the target pixel by using the linear
model of that group. If the luminance value of the target pixel is
greater than the average luminance value AL, the group G1 is
considered as being nearer to the target group, and the
chromaticity value of the target pixel is predicted by using the
linear model LM1 of the group G1. Conversely, if the luminance
value of the target pixel is smaller than the average luminance
value AL, the group G2 is considered as being nearer to the target
pixel, and the chromaticity value of the target pixel is predicted
by using the linear model LM2 of the group G2.
[0032] FIG. 5 shows a schematic diagram of grouping according to an
exemplary embodiment of the disclosure. Reference samples are
scanned, and it is determined whether each of the reference samples
belongs to any one existing group of the groups. The step of
grouping the reference samples includes: establishing a first group
according to a first reference sample of the reference samples, and
assigning the first reference sample to the first group;
determining whether to add a second reference sample of the
reference samples into the first group according to a sample
characteristic value of the second reference sample; if it is
determined to add the second reference sample into the second
group, updating a group characteristic value of the first group;
and if it determined not to add the second reference sample into
the second group, establishing a second group and calculating a
group characteristic value of the second group. The group
characteristic value includes any combination of: a position, a
representative luminance value, a representative chromaticity
component value, a maximum luminance, a minimum luminance, a
maximum chromaticity, a minimum chromaticity, and a reference
sample quantity. The sample characteristic value includes any
combination of: a luminance, at least one chromaticity component,
and a position. The details of the above are described in the
following examples.
[0033] For example, details of determining whether to add a second
reference sample of the reference samples into the first group
according to a sample characteristic value of the second reference
sample can be as described below. Two embodiments are given below;
that is, two sets of determination equations, and given that it is
determined that one or all reference samples Rn (where n is a
positive integer) satisfy any set of the determination equations,
it can be determined to add the reference sample Rn (where n is a
positive integer) into the group.
[0034] The first set of determination equations are: determining
whether a luminance value YRn and a chromaticity value CRn of the
reference sample Rn (where n is a positive integer) satisfy the
following equations to determine whether the reference sample Rn
belongs to the group Gi (where i=1 to A, and A represents the
quantity of existing groups) (if all of the four equations below
are satisfied, it is determined that the reference sample Rn
belongs to the group Gi; that is, if any one of the equations is
not satisfied, it is determined that the reference sample Rn does
not belong to the group Gi): YRn>Yi_group_min-Ymargin;
YRn<Yi_group_max+Ymargin; CRn>Ci_group_min-Cmargin; and
CRn<Ci_group_max+Cmargin.
[0035] In the above, Yi_group_min represents a minimum luminance
value in the group Gi, Ymargin and Cmargin respectively represent a
luminance range threshold (which may be an existing constant value)
and a chromaticity range threshold (which may be an existing
constant value), Yi_group_max represents a maximum luminance value
in the group Gi, Ci_group_min represents a minimum chromaticity
value in the group Gi, and Ci_group_max represents a maximum
chromaticity value in the group Gi.
[0036] The second set of determination equations are: determining
whether the luminance value YRn and the chromaticity value CRn of
the reference sample Rn satisfy the following equations to
determine whether the reference sample Rn belongs to the group Gi
(if all of the four equations below are satisfied, it is determined
that the reference sample Rn belongs to the group Gi; that is, if
any one of the equations is not satisfied, it is determined that
the reference sample Rn does not belong to the group Gi):
YRn>Yi_group_max-Ymargin; YRn<Yi_group_min+Ymargin;
CRn>Ci_group_max-Cmargin; and CRn<Ci_group_min+Cmargin.
[0037] If the reference sample Rn does not fall within any one
existing group, a new group is created and the reference sample Rn
is assigned to the new group.
[0038] After all of the reference samples are grouped, associated
linear models are respectively established for the groups.
[0039] A target group is selected from the groups according to the
luminance value of the target pixel, and the chromaticity value of
the target pixel is predicted by using a target linear model of the
target group. For example, assuming that the luminance value of the
target pixel falls between Yi_group_min and Yi_group_max, the group
Gi is selected as the target group, and the chromaticity value of
the target pixel is predicted by using the target linear model LMi
of the target group Gi.
[0040] Examples are described below. Initially after the first
reference sample is scanned, a first group G1 is established,
wherein the first group G1 currently includes only the first
reference sample.
[0041] Then, a second reference sample R2 is scanned, and it is
determined according to one of the two sets of equations above
whether the second reference sample R2 belongs to the first group
G1.
[0042] If it is determined that the second reference sample R2
belongs to the first group G1, the second reference sample R2 is
assigned to the first group G1, and Y1_group_min (selecting the
smaller between the luminance value of the first reference sample
and the luminance value of the second reference sample),
Y1_group_max (selecting the larger between the luminance value of
the first reference sample and the luminance value of the second
reference sample), C1_group_min (selecting the smaller between the
chromaticity value of the first reference sample and the
chromaticity value of the second reference value), and C1_group_max
(selecting the larger between the chromaticity value of the first
reference sample and the chromaticity value of the second reference
value) are accordingly updated.
[0043] Alternatively, if it is determined that the second reference
sample R2 does not belong to the first group G1, a new second group
G2 is established (similarly, the second group G2 currently
includes only the second reference sample R2). The above process is
repeated until all of the reference samples have been grouped.
[0044] That is, the grouping in FIG. 5 can be represented as: (1)
establishing a first group according to a first reference sample,
assigning the first reference sample to the first group, and
generating a minimum luminance value, a maximum luminance value, a
minimum chromaticity value and a maximum chromaticity value of the
first group, wherein the minimum luminance value is a luminance
value of the first reference sample, the maximum luminance value is
a luminance value of the first reference sample, the minimum
chromaticity value is a chromaticity value of the first reference
value, and the maximum chromaticity value is a chromaticity value
of the first reference sample; (2) determining according to a
luminance value and a chromaticity value of a second reference
value whether the second reference sample belongs to the first
group, with associated details as described above; (3) if the
second reference sample belongs to the first group, updating the
minimum luminance value of the first group as a minimum value
between the minimum luminance value of the first group and the
luminance value of the second reference sample, updating the
maximum luminance value of the first group as a maximum value
between the maximum luminance value of the first group and the
luminance value of the second reference sample, updating the
minimum chromaticity value of the first group as a minimum value
between the minimum chromaticity value of the first group and the
chromaticity value of the second reference sample, and updating the
maximum chromaticity value of the first group as a maximum value
between the maximum chromaticity value of the first group and the
chromaticity value of the second reference sample; (4) searching
all established groups to determine whether the second reference
sample belongs to any one of the groups; if the second reference
sample does not belong to at least one of the groups, generating a
second group, and generating a minimum luminance value, a maximum
luminance value, a minimum chromaticity value and a maximum
chromaticity value of the second group, wherein the minimum
luminance value of the second group is the luminance value of the
second reference sample, the maximum luminance value of the second
group is the luminance value of the second reference sample, the
minimum chromaticity value of the second group is the chromaticity
value of the second reference sample, and the maximum chromaticity
value of the second group is the chromaticity value of the second
reference sample.
[0045] In other possible exemplary embodiments of the present
application, the step of grouping the reference samples includes:
(1) establishing a first group according to a first reference
sample, assigning the first reference sample to the first group,
and generating a minimum luminance value and a maximum luminance
value of the first group, wherein the minimum luminance value of
the first group is a luminance value of the first reference sample,
and the maximum luminance value of the first group is the luminance
value of the first reference sample; (2) determining according to a
luminance value of a second reference sample whether the second
reference sample belongs to the first group, wherein it is
determined according to one of the two sets of equations below
whether the second reference sample belongs to the first group (if
any one set of equations are satisfied, the second reference sample
is assigned to the first group): (A) the first set of determination
equations: YRn>Yi_group_min-Ymargin; and
YRn<Yi_group_max+Ymargin; and (B) the second set of
determination equations: YRn>Yi_group_max-Ymargin; and
YRn<Yi_group_min+Ymargin; (3) if the second reference sample
belongs to the first group, updating the minimum luminance value of
the first group as a minimum value between the minimum luminance
value of the first group and the luminance value of the second
reference sample, and updating the maximum luminance value of the
first group as a maximum value between the maximum luminance value
of the first group and the luminance value of the second reference
sample; (4) searching the established groups to determine whether
the second reference sample belongs to any one of the groups; if
the second reference sample does not belong to at least one of the
groups, generating a second group, and generating a minimum
luminance value and a maximum luminance value of the second
group.
[0046] In other possible exemplary embodiments of the present
application, the step of grouping the reference samples includes:
(1) establishing a first group according to a first reference
sample, assigning the first reference sample to the first group,
and generating a minimum chromaticity value and a maximum
chromaticity value of the first group, wherein the minimum
chromaticity value of the first group is a chromaticity value of
the first reference sample, and the maximum chromaticity value of
the first group is the chromaticity value of the first reference
sample; (2) determining according to a chromaticity value of a
second reference sample whether the second reference sample belongs
to the first group, wherein it is determined according to the two
sets of equations below whether the second reference sample belongs
to the first group (if any one set of equations are satisfied, the
second reference sample is assigned to the first group): (A) the
first set of determination equations: CRn>Ci_group_min-Cmargin;
and CRn<Ci_group_max+Cmargin; and (B) the second set of
determination equations: CRn>Ci_group_max-Cmargin; and
CRn<Ci_group_min+Cmargin; (3) if the second reference sample
belongs to the first group, updating the minimum chromaticity value
of the first group as a minimum value between the minimum
chromaticity value of the first group and the chromaticity value of
the second reference sample, and updating the maximum chromaticity
value of the first group as a maximum value between the maximum
chromaticity value of the first group and the chromaticity value of
the second reference sample; (4) searching the established groups
to determine whether the second reference sample belongs to any one
of the groups; if the second reference sample does not belong to at
least one of the groups, generating a second group, and generating
a minimum chromaticity value and a maximum chromaticity value of
the second group.
[0047] Further, in an embodiment of the present application, a
reference block and a target block can be further segmented. FIG. 6
shows a schematic diagram of segmentation according to an exemplary
embodiment of the disclosure. A reference sample is scanned along
the X-axis and the Y-axis, and it is determined whether color
discontinuity occurs, wherein the determination is performed
according to: |C_R(i+1)-C_Ri|>Cthreshold and/or
|Y_R(i+1)-Y_Ri|>Ythreshold, wherein C_R(i+1) and C_Ri
respectively represent the chromaticity value of the reference
sample R(i+1) and the chromaticity value of the reference sample
Ri, Y_R(i+1) and Y_Ri respectively represent the luminance value of
the reference sample R(i+1) and the luminance value of the
reference sample R(i), and Cthreshold and Ythreshold respectively
represent a chromaticity threshold (a predetermined value) and a
luminance threshold (a predetermined value).
[0048] Further, in other exemplary embodiments of the present
application, the step of grouping the reference samples can
include: defining a quantity of the groups; establishing the groups
of a fixed quantity, and calculating individual group
characteristic values of the at least one group according to
individual sample characteristic values of a plurality of reference
samples included in the groups; and assigning the reference samples
into the at least one group according to the individual sample
characteristic values. The group characteristic value includes any
combination of: a position, a representative luminance value, a
representative chromaticity component value, a maximum luminance, a
minimum luminance, a maximum chromaticity, a minimum chromaticity,
and a reference sample quantity. The sample characteristic value
includes any combination of: a luminance, at least one chromaticity
component and a position.
[0049] After the spatial position of the color discontinuity is
identified, the reference block and the target block are segmented.
Taking FIG. 6 for example, after the color discontinuities D1 and
D2 are identified, the reference block RB1 is segmented into
reference sub-blocks RB1_1, RB1_2 and RB1_3. After the color
discontinuity D3 is identified, the reference block RB2 is
segmented into reference sub-blocks RB2_1 and RB2_2. Similarly,
according to the color discontinuities D1 to D3, the target block
TB is segmented into target sub-blocks TB_1, TB_2, TB_3, TB_4, TB_5
and TB_6. Then, the reference sub-blocks are grouped (with details
of the grouping as described above), and corresponding models are
established by using the grouped reference sub-blocks.
[0050] Alternatively, the reference sub-block RB1_1 and the
reference sub-block RB2_1 are referred when the chromaticity value
of the target sample of the target sub-block TB_1 is predicted, the
reference sub-block RB1_2 and the reference sub-block RB2_1 are
referred when the chromaticity value of the target sample of the
target sub-block TB_2 is predicted, the reference sub-block RB1_3
and the reference sub-block RB2_1 are referred when the
chromaticity value of the target sample of the target sub-block
TB_3 is predicted, the reference sub-block RB1_1 and the reference
sub-block RB2_2 are referred when the chromaticity value of the
target sample of the target sub-block TB_4 is predicted, the
reference sub-block RB1_2 and the reference sub-block RB2_2 are
referred when the chromaticity value of the target sample of the
target sub-block TB_5 is predicted, and the reference sub-block
RB1_3 and the reference sub-block RB2_2 are referred when the
chromaticity value of the target sample of the target sub-block
TB_6 is predicted. Details for predicting the chromaticity value
are as described previously, and are omitted herein. That is, when
any one of the target sub-blocks is predicted, at least one of the
reference sub-blocks is excluded.
[0051] A color discontinuity occurs, for example, when a white
reference sample is adjacent to a black reference sample, or when a
red reference sample is adjacent to a blue reference sample.
[0052] In the present application, a color discontinuity is
identified so as to segment a reference block and a target block.
When the chromaticity value of a target sub-block is predicted, a
reference sub-block having a more similar color is taken into
consideration, and a reference sub-block having a less similar
color is eliminated. That is, taking FIG. 6 for example, the color
of the target sub-block TB_1 is more similar to that of the
reference sub-block RB1_1 but less similar to that of the reference
sub-block RB1_2. Thus, to reconstruct the target sub-block TB_1,
the reference sub-block RB1_1 is taken into consideration while the
reference sub-block RB1_2 is excluded. Thus, the prediction result
can be more accurate.
[0053] That is to say, in one exemplary embodiment of the present
application, to perform segmentation, a spatial position having a
characteristic value discontinuity (e.g., a color discontinuity) is
identified, so as to segment the reference block and the target
block, that is, the reference block is segmented into at least one
reference sub-block, and a group is established for the at least
one reference sub-block; and a model is established by using the at
least one reference sub-block.
[0054] FIG. 7 shows a schematic diagram of segmentation according
to another exemplary embodiment of the disclosure. After the color
discontinuities D1 and D2 are identified, the reference block RB1
is segmented into reference sub-blocks RB1_1, RB1_2 and RB1_3;
after the color discontinuity D3 is identified, the reference block
RB2 is segmented into reference sub-blocks RB2_1 and RB2_2.
Similarly, according to the color discontinuities, the target block
TB is segmented into target sub-blocks TB_1, TB_2 and TB_3.
[0055] Next, the reference sub-block RB1_1 and the reference
sub-block RB2_1 are referred when the chromaticity value of the
target sample of the target block TB_1 is predicted, the reference
sub-block RB1_2 and the reference sub-block RB2_2 are referred when
the chromaticity value of the target sample of the target block
TB_2 is predicted, and the reference sub-block RB1_3 is referred
when the chromaticity value of the target sample of the target
block TB_3 is predicted. Details for predicting the chromaticity
value are as described previously, and are omitted herein.
[0056] Alternatively, in other exemplary embodiments of the present
application, the step of predicting the chromaticity value of the
target pixel includes: segmenting the target block into at least
one target sub-block according to a spatial position of a
characteristic value discontinuity, selecting the target group
according to the target sub-block to which the target pixel
belongs, and predicting the chromaticity value of the target pixel
by using the model of the target group.
[0057] FIG. 8 shows a schematic diagram of establishing a linear
model according to an exemplary embodiment of the disclosure. In
FIG. 8, a linear model is established by applying a linear
regression algorithm. Individual corresponding linear models are
established for the groups. The linear model established can be
represented as: Y=.alpha.x+.beta., where Y represents the
chromaticity value, x represents the luminance value, and .alpha.
and .beta. are as shown below:
.alpha. = N ( L ( n ) C ( n ) ) - L ( n ) C ( n ) N ( L ( n ) L ( n
) ) - L ( n ) L ( n ) ##EQU00001## .beta. = C ( n ) - .alpha. L ( n
) N ##EQU00001.2##
[0058] In the above, N represents the quantity of reference samples
in the group, L(n) and C(n) respectively represent the luminance
value and the chromaticity value of the reference sample Rn.
[0059] Alternatively, another straight line equation is:
Y=.alpha.x+.beta., where a and .beta. are as shown below:
.alpha. = C max - C min L max - L min , and ##EQU00002## .beta. = C
max - L max C max - C min L max - L min . ##EQU00002.2##
In the above, Cmax, Cmin, Lmax and Lmin respectively represent the
maximum chromaticity, the minimum chromaticity value, the minimum
luminance value and the maximum luminance value of the reference
samples.
[0060] FIG. 9A to FIG. 9C show schematic diagrams of establishing
linear models according to an exemplary embodiment of the
disclosure. In FIG. 9A to FIG. 9C, linear models are established by
applying straight line equations. That is, two points (two
reference samples) are identified from the reference samples to
establish a straight line, and the established straight line is a
linear model.
[0061] In FIG. 9A, the points identified are P1:(L1, Max C) and
P2:(L2, Min C); that is, P1 represents a reference sample having a
maximum chromaticity value among the reference samples in the
group, and P2 represents a reference sample having a minimum
chromaticity value among the reference samples in the group.
[0062] In FIG. 9B, the points identified are P1:(Max L1, C1) and
P2:(Min L, C2); that is, P1 represents a reference sample having a
maximum luminance value among the reference samples in the group,
and P2 represents a reference sample having a minimum luminance
value among the reference samples in the group.
[0063] In FIG. 9C, the points identified are P1:(Max L1, Max C) and
P2:(Min L, Min C); that is, P1 represents a reference sample having
both a maximum luminance value and a maximum chromaticity value
among the reference samples in the group, and P2 represents a
reference sample having both a minimum luminance value and a
minimum chromaticity value among the reference samples in the
group.
[0064] FIG. 10 shows a schematic diagram of establishing a linear
model according to an exemplary embodiment of the disclosure. In
FIG. 10, a linear model is established by applying an averaging
algorithm. In the application scenario of FIG. 10, the luminance
value distribution and/or chromaticity value distribution of the
reference samples in the group are quite concentrated, or the group
includes only one reference sample. When the averaging algorithm is
applied, the predicted chrominance value of the target sample is an
average value of the chromaticity values of the reference
samples.
[0065] FIG. 11 shows a schematic diagram of predicting a
chromaticity value according to an exemplary embodiment of the
disclosure. In an exemplary embodiment of the present application,
when the chromaticity value of the target pixel is predicted, a
group nearest to the target pixel is identified according to the
luminance value of the target pixel. Then, the chromaticity value
of the target pixel is predicted by applying the linear model of
the nearest group.
[0066] In other words, in FIG. 11, V1, V2 and V3 respectively
present average luminance values of the groups G1, G2 and G3. The
step of identifying a group nearest to the target pixel is:
determining whether the luminance value TP_Y of the target pixel is
nearest to V1, V2 or V3. If the luminance value TP_Y of the target
pixel is nearest to V1, it is determined that the group G1 is
nearest to the target pixel, and the chromaticity value TP_C of the
target pixel is predicted by applying the linear model LM1 of the
group G1. Similarly, if it is determined that the luminance value
TP_Y of the target pixel is nearest to V2, it is determined that
the group G2 is nearest to the target pixel, and the chromaticity
value TP_C of the target pixel is predicted by applying the linear
model LM2 of the group G2. If it is determined that the luminance
value TP_Y of the target pixel is nearest to V3, it is determined
that the group G3 is nearest to the target pixel, and the
chromaticity value TP_C of the target pixel is predicted by
applying the linear model LM3 of the group G3. In the example in
FIG. 11, the luminance value TP_C of the target pixel is nearest to
V2, and it is determined that the group G2 is nearest to the target
pixel and the chromaticity value TP_C of the target pixel is
predicted by applying the linear model LM2 of the group G2.
[0067] FIG. 12 shows a schematic diagram of predicting a
chromaticity value according to an exemplary embodiment of the
disclosure. The target pixel is regarded as an out-tier. In an
exemplary embodiment of the present application, if a target pixel
does not fall within a range of any group, the target pixel is
regarded as an out-tier. In one exemplary embodiment of the
disclosure, there are three approaches for predicting the
chromaticity value of a target pixel regarded as an out-tier.
[0068] In the first approach, the chromaticity value is predicted
by using a universal linear model (ULM). The "universal correlation
model" and/or "universal linear model" is obtained according to
individual center points of all of the groups.
[0069] In the second approach, an average value of predicted
adjacent chromaticity values is used as the predicted chromaticity
value of the target pixel regarded as an out-tier. The term "a
predicated adjacent chromaticity value" refers to a predetermined
chrominance of the remaining target pixels having similarly
luminance values.
[0070] In the third approach, an intermediate grayscale value is
used as the predicted chromaticity value of the target pixel
regarded as an out-tier. For example, assuming that the pixel value
is 10-bit, 512 is used as the predicted chromaticity value of the
target pixel regarded as an out-tier.
[0071] Further, in one exemplary embodiment of the present
application, if a chromaticity value of a target pixel previously
processed is near the luminance value of the target pixel, the
(predicted) chromaticity value of the target pixel previously
processed is used as the predicted chromaticity value of the target
pixel.
[0072] FIG. 13 shows a flowchart of a video encoding method
according to an exemplary embodiment of the disclosure. As shown in
FIG. 13, in step 1310, a plurality of reference samples are
collected. In step 1320, the reference samples are grouped to
generate at least one group. In step 1330, a model of the at least
one group is established. In step 1340, a target pixel is obtained
from a target block. In step 1350, a target group is selected from
the at least one group according to the target pixel. In step 1360,
a luminance value of the target pixel is introduced into a model of
the target group to predict a chromaticity value of the target
pixel. In step 1365, a coding value is generated, wherein the
coding value includes an index value.
[0073] FIG. 14 shows a function block diagram of a video decoder
according to an exemplary embodiment of the disclosure. The video
decoder 1400 according to an exemplary embodiment of the disclosure
includes a processor 1410, a memory 1420, a decoding module 1430
and an index receiving module 1440. The processor 1410, the memory
1420, the decoding module 1430 and the index receiving module 1440
are coupled to one another. The processor 1410 is for controlling
the video decoder 1400. The memory 1420 is for storing a reference
block and a target block. The decoding module 1430 can perform the
video decoding method above, and the index receiving module 1440
receives an index value from the video encoder. The decoding module
140 and the index receiving module 1440 can be implemented by
software executed by a processor, or be implemented by hardware
circuits.
[0074] FIG. 15 shows a function block diagram of a video encoder
according to an exemplary embodiment of the disclosure. The video
encoder 1500 according to an exemplary embodiment of the disclosure
includes a processor 1510, a memory 1520, an encoding module 1530
and an index selecting module 1540. The processor 1510, the memory
1520, the encoding module 1530 and the index selecting module 1540
are coupled to one another. The processor 1510 is for controlling
the video encoder 1500. The memory 1520 is for storing a reference
block and a target block. The encoding module 1530 can perform the
video encoding method described above. The index selecting module
1540 generates an index value from a coding value generated by the
coding module 1530. The encoding module 1530 and the index
selecting module 1540 can be implemented by software executed a
processor, or be implemented by hardware circuits.
[0075] As described above, the exemplary embodiments of the present
application effectively predict a chromaticity value. Thus, when
video data in a YCbCr format is encoded, only Y data needs to be
encoded, hence effectively reducing the bitrate of encoding. On the
other hand, CbCr data can be predicted by using the above method.
The exemplary embodiments of the present application are capable of
solving the issue of poor efficiency of independently encoding a
chromaticity value and enhancing the overall encoding efficiency.
Therefore, the present application can be applied to products
involving video compression related technologies, for example but
not limited to, webcams, digital cameras, digital video cameras,
handheld mobile devices and digital televisions.
[0076] It will be apparent to those skilled in the art that various
modifications and variations can be made to the disclosed
embodiments. It is intended that the specification and examples be
considered as exemplary only, with a true scope of the disclosure
being indicated by the following claims and their equivalents.
* * * * *