CN110418137A - A kind of rest block collection measured rate regulation method for intersecting subset guiding - Google Patents
A kind of rest block collection measured rate regulation method for intersecting subset guiding Download PDFInfo
- Publication number
- CN110418137A CN110418137A CN201910701685.5A CN201910701685A CN110418137A CN 110418137 A CN110418137 A CN 110418137A CN 201910701685 A CN201910701685 A CN 201910701685A CN 110418137 A CN110418137 A CN 110418137A
- Authority
- CN
- China
- Prior art keywords
- block
- observation
- subset
- rest
- collection
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
- H04N19/13—Adaptive entropy coding, e.g. adaptive variable length coding [AVLC] or context adaptive binary arithmetic coding [CABAC]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
- H04N19/17—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object
- H04N19/176—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object the region being a block, e.g. a macroblock
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
- H04N19/587—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving temporal sub-sampling or interpolation, e.g. decimation or subsequent interpolation of pictures in a video sequence
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Compression Of Band Width Or Redundancy In Fax (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
The present invention relates to a kind of rest block collection measured rates of intersection subset guiding to regulate and control method, comprising the following steps: target image is divided into several pieces, is classified to all pieces, each piece is included by the division of block collection and intersects subset or rest block collection；Sequentially block-by-block observation, and statistical observation vector is carried out to subset is intersected using initial measurement rate；It is that all kinds of optimization measured rates is arranged in rest block collection according to the statistical conditions for intersecting observation vector in subset, sequentially block-by-block is carried out to rest block collection using optimization measured rate and is observed, and hourly observation value；Each piece of observation carries out multi-direction prediction by spiral order block-by-block, executes quantization and the entropy coding of residual error.The present invention improves the total quality of image observation under the holding of average measurement rate.
Description
Technical field
The present invention relates to the observed parameter control technique fields in compressed sensing communication system, more particularly to a kind of intersection
The rest block collection measured rate of subset guiding regulates and controls method.
Background technique
Compressed sensing communication system can break through the limitation of nyquist sampling law, make to owe Nyquist rate sampling
Signal still can effectively restore, implementation complexity from measurement end to rebuild end transfer.Compressed sensing measurement end is from original
The observation data of low-dimensional are obtained in the high dimensional signal to begin, observation process is to carry out signal acquisition and data compression simultaneously.As one
The typical compression of images perception framework of kind, the perception of quantization splits' positions (QuantizedBlockCompressiveSensing,
QBCS) target image is divided into size is several pieces identical, every piece includes multiple contiguous pixels both horizontally and vertically, so
Sequentially block-by-block carries out independent observation afterwards, finally executes prediction, quantization and entropy coding.In QBCS measurement end, the rule of observing matrix
Mould depends on the size of block, does not increase with the increase of target image resolution ratio, reduces calculating and storage overhead.If desired right
The measured rate of block is adjusted, it is only necessary to the selection line number for changing observing matrix, to simplify hardware design.
It was found by the inventors of the present invention that QBCS measurement end carries out all pieces of target image using single-measurement rate at present
Observation, has ignored the feature difference of different zones, and the image-region for keeping information content little is also assigned same measured rate, difficult
To promote the observation quality of target image.Under the constraint of average measurement rate and single observation, how optimally to divide for all pieces
It is still a problem with measured rate.
Summary of the invention
Technical problem to be solved by the invention is to provide a kind of rest block collection measured rate regulation sides of intersection subset guiding
Method can improve the total quality of image observation under the holding of average measurement rate.
The technical solution adopted by the present invention to solve the technical problems is: providing a kind of rest block collection of intersection subset guiding
Measured rate regulates and controls method, comprising the following steps:
(1) target image is divided into several pieces, classified to all pieces, intersection is included by the division of block collection by each piece
Collection or rest block collection；
(2) sequentially block-by-block observation, and statistical observation vector is carried out to subset is intersected using initial measurement rate；
It (3) is that all kinds of optimization measured rates is arranged in rest block collection according to the statistical conditions for intersecting observation vector in subset,
It carries out sequentially block-by-block to rest block collection using optimization measured rate to observe, and hourly observation value；
(4) each piece of observation carries out multi-direction prediction by spiral order block-by-block, executes quantization and the entropy coding of residual error.
The step (1) specifically: target image is divided into several pieces, and by all pieces of secondary ordered pair of spiral from inside to outside
Layout serial number, and three kinds of block classifications: inner region, middle area and outskirt are divided into for all pieces according to the size of block serial number；From the 1st block
Start, selects a block to subset is intersected every K block, other pieces are included into rest block collection.
Specifically, QBCS measurement end prepares to obtain the target image that a width size is W × H pixel, wherein W is width,
H is height.Target image to be measured is divided into N=(W × H)/b2A block not overlapped, every block size are b × b picture
Element, N represent all pieces of target image of sum, and b represents the pixel number of every piece of side length.By all pieces of ordered pair of spiral from inside to outside time
Layout serial number, i expression block serial number, i=1,2, N is chosen as the 1st block positioned at the block of picture centre, enables xiIndicate the
The original pixels of i block, the block serial number being incremented by by spiral from inside to outside are divided into three kinds of block classifications for all pieces: inner region, middle area,
Outskirt, corresponding in target image, in, outer three regions, number of blocks is respectively IfIndicate that selection is not more than and immediate integer, j indicate block category label (j=1,2,3), then the block sequence of inner region (j=1)
Number from 1 toThe block serial number of middle area (j=2) fromIt arrivesThe block serial number of outskirt (j=3) fromTo N.From the 1st BOB(beginning of block), select a block to subset is intersected every K block, other pieces are included into rest block collection, this
Sample rest block integrates and intersects the ratio between number of blocks of subset as K-1.
The step (2) specifically: using initial measurement rate using the block of picture centre as starting point, by spiral order block-by-block into
Row independent observation, and cache observation.
Specifically, being based on initial measurement rate, intersects subset by spiral order block-by-block and execute observation, and cache observation.
Φ0It is a M0×b2The gaussian random observing matrix of size, whereinIt is the observation number for intersecting every piece of subset
Amount.Measurement end is from the BOB(beginning of block) of picture centre, by spiral order (block serial number i=1, K+1,2K+1) block-by-block to i-th
Block xiCarry out independent observation.Every piece uses observing matrix Φ0, i-th piece of xiObservation vector yiIt indicates are as follows: yi=
Φ0·xi, size M0×1.When intersecting subset execution block-by-block observation, observation counter obtains each piece of observation vector.
Intersecting subset and rest block collection has approximate statistical property, and measurement end is according to the statistical conditions for intersecting observation vector in subset
All kinds of optimization measured rates is set for rest block collection.
The step (3) specifically: according to the mean value weighting for the observation vector for intersecting subset different masses classification, for residue
Optimization measured rate is distributed in the region of block collection different masses classification respectively, wherein the distribution of biggish piece of classification of observation vector mean value compared with
High optimization measured rate, lesser piece of classification of observation vector mean value distribute lower optimization measured rate；According to different masses classification
Optimization measured rate, rest block collection observed by spiral order block-by-block, and caches observation.
After having observed all pieces that intersect subset, measured rate distributor is different masses class under the holding of average measurement rate
Do not distribute all kinds of optimization measured rates, calculate separately the mean value of three kinds of block classification observation vectors, obtain mean vector matrix delta=
[δ1,δ2,δ3], wherein δjFor intersect subset in jth kind block classification all observation vectors mean vector, j ∈ 1,2,
3}.According to inner region, middle area, three kinds of block classifications of outskirt mean value weighting, the corresponding region of every kind of block classification is pressedRedistribute optimization measured rate, in formula, | | * | |1For l1Normal form, based on
Calculate the sum of the absolute value of all elements in a vector, SminIt is allowed minimum measured rate.Based on the optimization measurement after regulation
Rate, rest block collection by spiral order (block serial number i=2, K, K+2,2K) block-by-block executes observation,
And observation is cached, ΦjIt is a Mj×b2The gaussian random observing matrix of size, whereinIt is jth kind block classification
Observation quantity.Belong to i-th piece of x of jth kind block classificationiObservation vector indicate are as follows: yi=Φj·xi, wherein yiIt is
I-th of observation vector, size Mj×1。
The step (4) specifically: multi-direction prediction is executed to observation from central block by spiral order block-by-block, from
Available adjacent block is chosen in the multi-direction neighborhood of current block, the observation prediction for calculating separately each adjacent block and current block is residual
Difference selects the smallest prediction direction of residual error and optimum prediction residual error, executes quantization to the optimum prediction residual error of current block and compiles with entropy
Code, generates the code stream of storage or transmission.
Beneficial effect
Due to the adoption of the above technical solution, compared with prior art, the present invention having the following advantages that and actively imitating
Fruit: the present invention is using measured rate as regulation parameter, and by observing by spiral order block-by-block, statistics intersection three kinds of block classifications of subset are seen
The mean value weighting of measured value vector, for rest block collection three kinds of block uneven class sizes distribute optimization measured rate, adaptively improve
The utilization efficiency of measured rate.The present invention can effectively distribute measured rate under the holding of average measurement rate, mention to a certain extent
The high observation quality of target image.
Detailed description of the invention
Fig. 1 is the entire block diagram of QBCS measurement end；
Fig. 2 is the intersection subset and rest block collection schematic diagram of sequentially block-by-block observation；
Fig. 3 is the rest block collection measured rate regulation method flow diagram for intersecting subset guiding.
Specific embodiment
Present invention will be further explained below with reference to specific examples.It should be understood that these embodiments are merely to illustrate the present invention
Rather than it limits the scope of the invention.In addition, it should also be understood that, after reading the content taught by the present invention, those skilled in the art
Member can make various changes or modifications the present invention, and such equivalent forms equally fall within the application the appended claims and limited
Range.
Embodiments of the present invention are related to a kind of rest block collection measured rate regulation method of intersection subset guiding, including following
Step: being divided into several pieces for target image, classify to all pieces, and each piece is included by the division of block collection and intersects subset or surplus
Remaining block collection；Sequentially block-by-block observation, and statistical observation vector is carried out to subset is intersected using initial measurement rate；According to intersection subset
The statistical conditions of middle observation vector are that all kinds of optimization measured rates is arranged in rest block collection, using optimization measured rate to rest block collection
It carries out sequentially block-by-block to observe, and hourly observation value；Each piece of observation carries out multi-direction prediction by spiral order block-by-block, executes residual
The quantization of difference and entropy coding.It can be seen that present embodiment is observed by the block-by-block for first executing initial measurement rate to intersection subset,
The optimization measured rate of various pieces of classifications of rest block collection is derived, with this to improve observation quality under the holding of average measurement rate.
Embodiments of the present invention that the following is further explained with reference to the attached drawings.In the present embodiment, the target image X of input is big
Small is 160 × 160 pixels, divides the blocking complexity for being observed, helping to reduce measurement end.Fig. 1 gives QBCS measurement
The entire block diagram at end is mainly divided comprising image block/sub-category, block collection, sequentially block-by-block observation, observation counter, measurement
The functional units such as rate distributor, quantization, entropy coding.Measurement end be first carried out target image X piecemeal and all pieces of classification
Not, intersection subset or rest block collection are included by the division of block collection each piece.Based on certain initial measurement rate, measurement end is sub to intersecting
Collection sequentially observe by block-by-block, and observation counter obtains the observation vector of various pieces of classifications.Higher measured rate normally results in
Better observation quality, the biggish image-region of information content should be assigned higher optimization measured rate.Intersect subset and rest block
Various pieces of classifications of collection have similar provincial characteristics.Measured rate distributor obtain intersect various pieces of classifications of subset observation to
The mean value of amount is rest block collection different masses classification distribution optimization measured rate according to mean value weighting.Under the holding of average measurement rate, survey
Measuring end, to rest block collection, sequentially block-by-block is observed according to the optimization measured rate of various pieces of classifications.The observation of real number carries out multi-direction pre-
It surveys, residual quantization becomes index value, and then executes entropy coding, and compressed data and auxiliary information are encapsulated into code stream Y, carry out number
The storage or transmission of change.
The block serial number being incremented by by spiral is needed to be divided into three kinds of block classifications for all pieces in the present embodiment: inner region, Zhong Qu, outer
Area.Mentioned method first observes intersection subset and observes rest block collection again, and the sequentially block-by-block observation of two set of blocks is by by introversion
The order that outer piece of serial number increases is observed.Fig. 2 gives the intersection subset (black block composition) and residue of sequentially block-by-block observation
Block collection (white blocks composition) schematic diagram shows that the target image X containing N=100 block carries out spiral block-by-block observation in figure
An example, the spiral order of the digital representation block in figure in each box, the number in box is smaller, indicates that the block is got over
It is early to execute observation, intersect subset and rest block collection is observed by the order block-by-block that block serial number from inside to outside increases.
Fig. 3 gives the rest block collection measured rate regulation method flow diagram for intersecting subset guiding, and measurement end execution first intersects
The sequentially block-by-block observation of subset, rear rest block collection, generates code stream, detailed process is as follows:
Step 1: carrying out piecemeal to target image, be then all pieces of divided block classifications, and each piece is included by the division of block collection
Intersect subset or rest block collection.QBCS measurement end is every piece big first by the target image X of the input points of blocks for 100 non-overlaps
Small is 16 × 16 pixels, and by spiral all pieces of layout serial numbers of time ordered pair, i indicates block serial number, i=1,2,100,Indicate i-th piece of original pixels.Block xiBy gaussian random observing matrix, observation vector y is obtainedi.1st
Block (i=1) is located at the right of the 1st block from the BOB(beginning of block) of picture centre, the 2nd block, and the 3rd block is located at the lower section of the 2nd block,
The rest may be inferred.100 blocks are divided into three kinds of block classifications by the block serial number that measurement end is incremented by by spiral from inside to outside: inner region, middle area,
Outskirt, correspond respectively to target image X it is interior, in, outer three regions, the number of blocks of relevant block classification is 33,33,34 respectively.
If j indicate block category label (j=1,2,3), then the block serial number of inner region (j=1) from 1 to 33, the block serial number of middle area (j=2) from
34 to 66, the block serial number of outskirt (j=3) is from 67 to 100.From the 1st BOB(beginning of block), a block is selected to be included into friendship every K=3 block
Fork collection, other pieces are included into rest block collection, and such rest block collection and the ratio between the number of blocks for intersecting subset are approximately 2:1.
Step 2: intersect the sequentially block-by-block observation of subset, statistical observation vector.All pieces for intersecting subset are all made of just
Beginning measured rate S0, Φ0It is a M0The gaussian random observing matrix of × 256 sizes, whereinIt is to intersect subset
Each piece of observation quantity,Expression selection is not more than and immediate integer.Based on initial measurement rate S0, from image
The BOB(beginning of block) of the heart intersects subset by spiral order (block serial number i=1,4,7) block-by-block to i-th piece of xiIt carries out independent
Observation, every piece uses gaussian random observing matrix Φ0, i-th piece of xiCorresponding observation vector yiIt indicates are as follows: yi=Φ0·
xi, size M0× 1, block-by-block caches observation.When intersecting subset and executing block-by-block observation, in observation counter obtains respectively
Area, middle area, three kinds of block classifications of outskirt observation vector.Have relatively uniform statistics special due to intersecting subset and rest block collection
Property, measurement end will be rest block collection setting optimization measured rate according to the mean value weighting for intersecting observation vector in subset.
Step 3: optimize the distribution of measured rate, the sequentially block-by-block observation of rest block collection.Measurement end is kept in average measurement rate
Under, respective optimization measured rate is distributed for different masses classification.Measured rate distributor, which calculates separately, intersects subset inner region, Zhong Qu, outer
The observation vector of the three kinds of block classifications in area, thus obtains mean vector matrix delta=[δ1,δ2,δ3], wherein δjTo intersect in subset
The mean vector of all observation vectors of jth kind block classification, j ∈ { 1,2,3 }.Added according to the mean value of inner region, middle area, outskirt
Power, measured rate distributor are pressedJth kind block classification is concentrated again for rest block
It distributes and optimizes measured rate, in formula, | | * | |1For l1Normal form, for calculating the sum of the absolute value of all elements in a mean vector,
The minimum measured rate S of permissionmin=0.01, max (*, *) indicate selection maximum value therein, and min (*, *) indicates to choose therein
Minimum value.Rest block collection is used by the observation of spiral order (block serial number i=2,3,5,6,8,9) block-by-block, various pieces of classifications
Respective optimization measured rate Sj, corresponding ΦjIt is a MjThe gaussian random observing matrix of × 256 sizes, whereinIt is the observation quantity of jth kind block classification.Belong to i-th piece of x of jth kind block classificationiObservation export indicate
Are as follows: yi=Φj·xi, wherein yiIt is i-th of observation vector, size Mj× 1, block-by-block caches observation.
Step 4: the coding of observation.Adjacent block has stronger correlation in image, and adjacent block also has in observation domain
There is stronger correlation, multi-direction prediction can eliminate the redundancy between observation.Measurement end is by spiral order block-by-block to current
The observation of block executes multi-direction prediction, chooses available adjacent block from the multi-direction neighborhood of current block, calculates separately each
The observation prediction residual of adjacent block and current block therefrom selects the smallest prediction direction of residual error and optimum prediction residual error.Currently
The optimum prediction residual error of block uses vector quantization, and the data after quantization execute arithmetic entropy coding, by compressed data and auxiliary information
It is encapsulated into code stream Y, carries out storage or transmission.End is rebuild after receiving code stream Y, restores to obtain using typical image reconstruction algorithm
Reconstruction image.Under the holding of average measurement rate, mentioned method can improve the observation quality of target image X to a certain extent, weight
Image is built to obtain higher Y-PSNR.
Claims (6)
1. a kind of rest block collection measured rate for intersecting subset guiding regulates and controls method, which comprises the following steps:
(1) target image is divided into several pieces, classified to all pieces, by each piece by block collection division be included into intersect subset or
Rest block collection；
(2) sequentially block-by-block observation, and statistical observation vector is carried out to subset is intersected using initial measurement rate；
(3) it is that all kinds of optimization measured rates is arranged in rest block collection according to the statistical conditions for intersecting observation vector in subset, uses
Optimize measured rate and sequentially block-by-block observation, and hourly observation value is carried out to rest block collection；
(4) each piece of observation carries out multi-direction prediction by spiral order block-by-block, executes quantization and the entropy coding of residual error.
2. the rest block collection measured rate according to claim 1 for intersecting subset guiding regulates and controls method, which is characterized in that described
Step (1) specifically: target image is divided into several pieces, and by all pieces of layout serial numbers of secondary ordered pair of spiral from inside to outside, and
Three kinds of block classifications: inner region, middle area and outskirt are divided by all pieces according to the size of block serial number；From the 1st BOB(beginning of block), every K
A block selects a block to subset is intersected, and other pieces are included into rest block collection.
3. the rest block collection measured rate according to claim 1 for intersecting subset guiding regulates and controls method, which is characterized in that described
Step (2) specifically: using initial measurement rate using the block of picture centre as starting point, independent observation is carried out by spiral order block-by-block,
And cache observation.
4. the rest block collection measured rate according to claim 1 for intersecting subset guiding regulates and controls method, which is characterized in that described
Step (3) specifically: be rest block collection different masses class according to the mean value weighting for the observation vector for intersecting subset different masses classification
Optimization measured rate is distributed in other region respectively, wherein biggish piece of classification of observation vector mean value distributes higher optimization measurement
Rate, lesser piece of classification of observation vector mean value distribute lower optimization measured rate；According to the optimization measured rate of different masses classification,
Rest block collection is observed by spiral order block-by-block, and caches observation.
5. the rest block collection measured rate according to claim 4 for intersecting subset guiding regulates and controls method, which is characterized in that described
Distribution optimization measured rate when, the corresponding region of every kind of block classification according toIts
In, | | * | |1For l1Normal form, for calculating the sum of the absolute value of all elements in a vector；δjTo intersect jth kind block in subset
The mean vector of all observation vectors of classification, SminFor the minimum measured rate of permission, S0It is expressed as initial measurement rate.
6. the rest block collection measured rate according to claim 1 for intersecting subset guiding regulates and controls method, which is characterized in that described
Step (4) specifically: multi-direction prediction is executed to observation from central block by spiral order block-by-block, from the multi-party of current block
Available adjacent block is chosen into neighborhood, calculates separately the observation prediction residual of each adjacent block and current block, selects residual error
The smallest prediction direction and optimum prediction residual error execute quantization and entropy coding to the optimum prediction residual error of current block, generate storage
Or the code stream of transmission.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910701685.5A CN110418137B (en) | 2019-07-31 | 2019-07-31 | Cross subset guided residual block set measurement rate regulation and control method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910701685.5A CN110418137B (en) | 2019-07-31 | 2019-07-31 | Cross subset guided residual block set measurement rate regulation and control method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110418137A true CN110418137A (en) | 2019-11-05 |
| CN110418137B CN110418137B (en) | 2021-05-25 |
Family
ID=68364825
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910701685.5A Active CN110418137B (en) | 2019-07-31 | 2019-07-31 | Cross subset guided residual block set measurement rate regulation and control method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110418137B (en) |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2006122146A2 (en) * | 2005-05-10 | 2006-11-16 | William Marsh Rice University | Method and apparatus for distributed compressed sensing |
| US20100080473A1 (en) * | 2008-09-26 | 2010-04-01 | Futurewei Technologies, Inc. | System and Method for Compressing and Decompressing Images and Video |
| WO2012044380A2 (en) * | 2010-06-18 | 2012-04-05 | William Marsh Rice University | Method and apparatus for compressive acquisition and recovery of dynamic imagery |
| CN105096345A (en) * | 2015-09-15 | 2015-11-25 | 电子科技大学 | Target tracking method based on dynamic measurement matrix and target tracking system based on dynamic measurement matrix |
| CN106056638A (en) * | 2016-05-18 | 2016-10-26 | 信阳师范学院 | Low-complexity compressive sensing image coding method adapted to texture contrast |
| CN106559670A (en) * | 2016-11-07 | 2017-04-05 | 湖南源信光电科技有限公司 | A kind of improved piecemeal video compress perception algorithm |
| CN106941609A (en) * | 2017-02-15 | 2017-07-11 | 浙江工业大学 | The video-frequency compression method perceived based on self adaptation splits' positions |
| CN108235020A (en) * | 2018-01-05 | 2018-06-29 | 东华大学 | A kind of spiral block-by-block measured value Forecasting Methodology perceived towards quantization splits' positions |
| CN109448065A (en) * | 2018-10-11 | 2019-03-08 | 北京理工大学 | A kind of compression sensing method based on the measurement of gradient block adaptive |
| CN109819258A (en) * | 2019-01-30 | 2019-05-28 | 东华大学 | A kind of splits' positions perceived direction predictive coding method based on spiral scan |
-
2019
- 2019-07-31 CN CN201910701685.5A patent/CN110418137B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2006122146A2 (en) * | 2005-05-10 | 2006-11-16 | William Marsh Rice University | Method and apparatus for distributed compressed sensing |
| US20100080473A1 (en) * | 2008-09-26 | 2010-04-01 | Futurewei Technologies, Inc. | System and Method for Compressing and Decompressing Images and Video |
| WO2012044380A2 (en) * | 2010-06-18 | 2012-04-05 | William Marsh Rice University | Method and apparatus for compressive acquisition and recovery of dynamic imagery |
| CN105096345A (en) * | 2015-09-15 | 2015-11-25 | 电子科技大学 | Target tracking method based on dynamic measurement matrix and target tracking system based on dynamic measurement matrix |
| CN106056638A (en) * | 2016-05-18 | 2016-10-26 | 信阳师范学院 | Low-complexity compressive sensing image coding method adapted to texture contrast |
| CN106559670A (en) * | 2016-11-07 | 2017-04-05 | 湖南源信光电科技有限公司 | A kind of improved piecemeal video compress perception algorithm |
| CN106941609A (en) * | 2017-02-15 | 2017-07-11 | 浙江工业大学 | The video-frequency compression method perceived based on self adaptation splits' positions |
| CN108235020A (en) * | 2018-01-05 | 2018-06-29 | 东华大学 | A kind of spiral block-by-block measured value Forecasting Methodology perceived towards quantization splits' positions |
| CN109448065A (en) * | 2018-10-11 | 2019-03-08 | 北京理工大学 | A kind of compression sensing method based on the measurement of gradient block adaptive |
| CN109819258A (en) * | 2019-01-30 | 2019-05-28 | 东华大学 | A kind of splits' positions perceived direction predictive coding method based on spiral scan |
Non-Patent Citations (2)
| Title |
|---|
| 李然，干宗良，崔子冠等: "图像分块压缩感知中的自适应测量率审定方法", 《通信学报》 * |
| 潘菲菲: "基于图像块分类与自适应多字典学习的图像压缩感知", 《电脑知识与技术》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110418137B (en) | 2021-05-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109829855B (en) | Super-resolution reconstruction method based on fusion of multi-level feature images | |
| CN104023235B (en) | Indexed color history in Image Coding | |
| CN104221383B (en) | Picture coding device, picture decoding apparatus, image encoding method and picture decoding method | |
| CN105872565B (en) | Picture coding device and method and image decoder and method | |
| CN110097609B (en) | Sample domain-based refined embroidery texture migration method | |
| CN100459707C (en) | Mode coding method and apparatus for use in interlaced shape coder | |
| CN108681991A (en) | Based on the high dynamic range negative tone mapping method and system for generating confrontation network | |
| CN109672451A (en) | Demura data compression method, mura compensation method and device | |
| CN109214971B (en) | A kind of gray level image visual encryption method | |
| CN107454405A (en) | Method for displaying image including method for encoding images and picture decoding method | |
| CN110139102B (en) | Method, device, equipment and storage medium for predicting video coding complexity | |
| CN109919840A (en) | Image super-resolution rebuilding method based on dense feature converged network | |
| CN104143186B (en) | A kind of SLIC super-pixel piecemeal optimization method | |
| CN106941609B (en) | Video-frequency compression method based on the perception of adaptive splits' positions | |
| CN108923984A (en) | Space-time video compress cognitive method based on convolutional network | |
| CN106030652A (en) | Method of and apparatus for displaying an output surface in data processing systems | |
| CN111754928A (en) | Display screen layered correction method and device | |
| CN100515096C (en) | Method and device for digital image adaptative color adjustment of display | |
| CN107087171A (en) | HEVC integer pixel motion estimation methods and device | |
| CN113052859A (en) | Super-pixel segmentation method based on self-adaptive seed point density clustering | |
| CN101710985A (en) | Image brightness compensation method for image coding | |
| CN110418137A (en) | A kind of rest block collection measured rate regulation method for intersecting subset guiding | |
| Chen et al. | Underwater-image super-resolution via range-dependency learning of multiscale features | |
| CN101742320A (en) | Image processing method | |
| CN109618156A (en) | A kind of video frequency coding rate method of adjustment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |