CN103473797B - Spatial domain based on compressed sensing sampling data correction can downscaled images reconstructing method - Google Patents
Spatial domain based on compressed sensing sampling data correction can downscaled images reconstructing method Download PDFInfo
- Publication number
- CN103473797B CN103473797B CN201310421657.0A CN201310421657A CN103473797B CN 103473797 B CN103473797 B CN 103473797B CN 201310421657 A CN201310421657 A CN 201310421657A CN 103473797 B CN103473797 B CN 103473797B
- Authority
- CN
- China
- Prior art keywords
- image
- sampling
- image block
- matrix
- designated
- 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.)
- Expired - Fee Related
Links
Landscapes
- Compression Or Coding Systems Of Tv Signals (AREA)
- Compression Of Band Width Or Redundancy In Fax (AREA)
Abstract
The invention discloses a kind of spatial domain based on compressed sensing sampling data correction can downscaled images reconstructing method, it is the quality problems that can reduce the image finally produced for the existing method based on utilizing " reconstruct-spatial domain, compressed sensing territory down-sampling " two steps to realize low-resolution image displaying in the method for compressing image of compressed sensing, by combining reconstruct can be reduced to the compressed sensing sampling of image with spatial domain, realize the object of directly structure high-quality low-resolution image under the condition of low sampling rate.Compared with reconstructing method can being reduced with the existing spatial domain based on compressed sensing, the present invention has the advantages that to provide high-quality low-resolution image efficiently, the spatial domain using " reconstruct-spatial domain, compressed sensing territory down-sampling " two steps to realize image in existing method can be overcome and can reduce reconstruct, and high-quality reference picture cannot be provided, and then the shortcoming of high-quality low-resolution image cannot be produced.
Description
Technical field
The invention belongs to image processing field, the spatial domain relating generally to digital picture reduces reconfiguration technique.
Background technology
Image derives from natural scene usually, the shooting recording process of any image, and essence is the process of original scene being carried out data sampling and compression.Along with the development of network technology and Emerging multimedia technology, require that image procossing not only will adapt to network environment complicated and changeable, and changeable display environment will be adapted to.How realizing efficient multiresolution image processing, the multiresolution image processing particularly under low sampling rate condition, to adapt to changeable display environment, is a major challenge that current image processing field faces.And the low resolution display how realized under emerging compress technique original scene or image, be the problem needing to solve of standing in the breach.
For a long time, in order to can intactly reconstruction signal, traditional sampling process all will be followed nyquist sampling theorem to realize.This sampling theory has been used 80 years nearly in signal transacting field, until the proposition of compressive sensing theory, just broken this sampling criterion.2006, Donobo, the people such as Candes and Tao propose the compressive sensing theory for sparse signal process, according to compressive sensing theory, can be required in far below nyquist sampling theorem sample frequency stochastic sampling is carried out to signal, thus obtain a small amount of observed reading, then by solving convex optimization problem, these observed readings can be utilized to carry out Accurate Reconstruction to original signal, to realize the compression of data.An important application of this theory carries out compression process to image exactly, and this is also an important breakthrough of image processing field in recent years.To sample the data obtained through compressed sensing, while completing raw data sampling, also complete the compression to data, this is an efficient information process inherently.
When utilizing the existing compression of images based on compressed sensing and reconfiguration technique to carry out low-resolution image reconstruct to the data after over-sampling, need to realize in two steps: the first step, the data after sampling are carried out and the reconstruct of original image with resolution in compressed sensing territory; Second step, carries out down-sampling to the image after reconstruct in spatial domain, obtains the image of low resolution.The method that this substep realizes, there is a great defect: when sampling rate is lower, the same image in different resolution reconstructed is second-rate, and recycle the image that this low-quality image goes to produce low resolution, the quality of the image finally obtained also can be poor.In order to overcome this defect, the data that patent utilization compression sampling of the present invention obtains directly construct high-quality low-resolution image in compressed sensing territory, to realize reducible Image Reconstruction.
Summary of the invention
The object of this invention is to provide a kind of can the reconstructing method of downscaled images based on compressed sensing sampling spatial domain, it is by combining reducing reconstruct to the compressed sensing sampling of image with spatial domain, realizing the object of directly structure high-quality low-resolution image under the condition of low sampling rate.Compared with reconstructing method can being reduced with the existing spatial domain based on compressed sensing, the present invention has the advantages that to provide high-quality low-resolution image efficiently, the spatial domain using " reconstruct-spatial domain, compressed sensing territory down-sampling " two steps to realize image in existing method can be overcome and can reduce reconstruct, and high-quality reference picture cannot be provided, and then the shortcoming of high-quality low-resolution image cannot be produced.
Content of the present invention for convenience of description, first do following term definition:
Definition 1, traditional image block method
Traditional image block method is according to the method for in Joint Photographic Experts Group, image being carried out to piecemeal, original image is divided into the equidimension image block of multiple non-overlapping copies, specific descriptions process is see " JPEG (JointPhotographicExpertsGroup): ISO/IECIS10918 – 1/ITU-TRecommendationT.81; DigitalCompressionandCodingofContinuous-ToneStillImage, 1993 ";
Definition 2, the method for traditional image block composograph
The method of traditional image block composograph carries out not overlapping each other combination with the method for synthesizing complete image with image block according in Joint Photographic Experts Group, specific descriptions process is see " JPEG (JointPhotographicExpertsGroup): ISO/IECIS10918 – 1/ITU-TRecommendationT.81; DigitalCompressionandCodingofContinuous-ToneStillImage, 1993 ";
Definition 3, traditional compressed sensing sampling
Traditional compressed sensing method of sampling is under the condition lower than nyquist sampling rate, according to the sampling rate of setting, produce stochastic sampling matrix Ф, with stochastic sampling matrix Ф, original signal is sampled, obtain the method for the discrete sample of signal, the process that specifically describes is see document " Compressedsensing ".
Definition 4, traditional compressed sensing based on image block is sampled
In order to realize sampling to the compressed sensing of two dimensional image, usually first by two dimensional image piecemeal, again each two dimensional image block is converted into an one-dimensional signal, final stochastic sampling matrix Ф carries out compression sampling to each one-dimensional signal, and concrete steps are see document " Blockcompressedsensingofnaturalimages ".
Definition 5, the sparse reconstruction method of traditional gradient projection
Sparse reconstruction method (the Gradientprojectionforsparsereconstruction of gradient projection, GPSR) be that convex optimization problem is converted into quadratic programming problem by one, and utilize special line search method and termination tech, data after compressed sensing sampling are carried out to the method for quick reconfiguration, concrete steps are see document " GradientProjectionforSparseReconstruction:ApplicationtoC ompressedSensingandOtherInverseProblems ".
Definition 6, the image block based on compressed sensing is rebuild
It is after carrying out the compression sampling based on image block to image that image block based on compressed sensing is rebuild, the problem that the data separate obtained sampling solves convex optimization is rebuild, produce the reconstruction signal of one dimension, again the reconstruction signal of one dimension is converted into the picture signal of two dimension, be used for realizing the reconstruction of whole image, concrete steps are see document " Blockcompressedsensingofnaturalimages ".
Definition 7, traditional image is at equal intervals without weighting spatial domain down-sampling method
Traditional image at equal intervals spatial domain is used to without weighting down-sampling method a kind of method realizing image down display.This method with first, image upper left corner pixel for starting point, simultaneously horizontally right and vertically downward both direction directly do not repeat at equal intervals to extract to pixel, and do not do the method for sampling of any neighborhood weighting process.
Definition 8, traditional method of matrix being carried out to submatrix division
Traditional carries out the method for submatrix division according to the method for in linear algebra theory, matrix being carried out to piecemeal to matrix, original matrix is divided into the equidimension submatrix of multiple non-overlapping copies, specific descriptions process is see " engineering mathematics linear algebra (the 5th edition) ", department of mathematics of Tongji University writes, Higher Education Publishing House.
Definition 9, Matlab
Matlab is the abbreviation of matrix labotstory (MatrixLaboratory), the business mathematics software of U.S. MathWorks Company, for advanced techniques computational language and the interactive environment of algorithm development, data visualization, data analysis and numerical evaluation.Detailed directions refers to document " the practical study course (second edition) of MATLAB ", and HollyMoore writes, and Gao Huisheng, Liu Tongna, Li Congcong translate, Electronic Industry Press.
The invention provides a kind of spatial domain based on compressed sensing sampling can the reconstructing method of downscaled images, and it comprises the following steps, as shown in Figure 1:
Step 1, the pre-service of original image
Be the original image of W × H by resolution, image block method is traditionally divided into N=(W × H)/m
2individual non-overlapping copies, size is the image block of m × m, is designated as B
0, B
1..., B
n-1, W represents the width of original image here, and H represents the height of original image, and N represents the number of original picture block, and m represents width and the height of original picture block.
Step 2, the compressed sensing sampling of image
To each image block B produced in step 1
0, B
1..., B
n-1realize traditional compressed sensing based on image block one by one with stochastic sampling matrix Ф to sample, obtain sampled data G
0, G
1..., G
n-1.Pooled sampling data G
0, G
1..., G
n-1, be designated as G, G={G
0, G
1..., G
n-1, G is the compressed sensing sampled data of original image.Here, the number of N representative image block, the columns of stochastic sampling matrix Ф is m
2, line number is determined by the sampling rate in the sampling of traditional compressed sensing.
Step 3, waits the reconstruct of image in different resolution
Adopt traditional sparse reconstruction method of gradient projection to each the sampled data G obtained in step 2
0, G
1..., G
n-1realize traditional image block based on compressed sensing one by one to rebuild, obtain rebuilding the image block that rear size is n × n, be designated as
,
will by the method for traditional image block composograph
form complete reconstruction image, be designated as I, the resolution of rebuilding image I is W × H, and here, the width of image I is rebuild in W representative, and the height of image I is rebuild in H representative; The number of image block is rebuild in N representative; Width and the height of image block are rebuild in n representative, and n=m, m represent width and the height of the original picture block produced in step 1.
Step 4, carries out piecemeal to reconstruction image
First, to the reconstruction image I produced in step 3, image is traditionally at equal intervals without weighting spatial domain down-sampling method, and setting sampling interval distance, is designated as L, L=2
d< m, d are positive integer, and m represents width and the height of the original picture block produced in step 1.Then, reconstruction image I image block method is traditionally divided into s=(W × H)/l
2individual non-overlapping copies, size is the image block of l × l, is designated as b
0, b
1..., b
s-1, and by image block b
iin pixel be designated as b
i(x, y); Here, the width of image I is rebuild in W representative, and the height of image I is rebuild in H representative; Width and the height of image block are rebuild in l representative, l=L; The index subscript of image block is rebuild in i representative, i=0, and 1 ..., s-1; The horizontal ordinate of pixel in image block is rebuild in x representative, and the ordinate of pixel in image block is rebuild in y representative, 0≤x<l, 0≤y<l.
Step 5, the generation of residual image
First, defining a resolution is the residual image of W × H, is designated as R, and the pixel value of all pixels in R is set as 0; Residual image R image block method is traditionally divided into s=(W × H)/l
2individual non-overlapping copies, size is the image block of l × l, is designated as r
0, r
1..., r
s-1, by image block r
i(i=0,1 ..., s-1) in pixel be designated as r
i(x, y) (0≤x<l, 0≤y<l); Here, W represents the width of residual image R, and H represents the height of residual image R; L represents width and the height of residual image block, and l=L, L represent the image that sets in step 4 at equal intervals without the sampling interval distance in the down-sampling of weighting spatial domain; I represents the index subscript of residual image block; X represents the horizontal ordinate of pixel in residual image block, and y represents the ordinate of pixel in residual image block;
Secondly, image block b is rebuild with s that produces in step 4
i(i=0,1 ..., s-1), the residual image block r of s correspondence is set
i(i=0,1 ..., s-1) in the pixel value of pixel, step is as follows:
To the 1st residual image block r
0, make r
0(0,0)=b
0(0,0)-b
0(0,0), r
0(0,1)=b
0(0,0)-b
0(0,1) ..., r
0(l-1, l-1)=b
0(0,0)-b
0(l-1, l-1); To the 2nd residual image block r
1, make r
1(0,0)=b
1(0,0)-b
1(0,0), r
1(0,1)=b
1(0,0)-b
1(0,1) ..., r
1(l-1, l-1)=b
1(0,0)-b
1(l-1, l-1); In like manner, to i residual image block, r is made
i-1(0,0)=b
i-1(0,0)-b
i-1(0,0), r
i-1(0,1)=b
i-1(0,0)-b
i-1(0,1) ..., r
i-1(l-1, l-1)=b
i-1(0,0)-b
i-1(l-1, l-1); Here, the index subscript of i representative image block; The width of l representative image block and height, l=L, L represent the image that sets in step 4 at equal intervals without the sampling interval distance in the down-sampling of weighting spatial domain; The number of behalf image block, s=(W × H)/l
2, W represents the width of residual image R, and H represents the height of residual image R;
Finally, by the method for traditional image block composograph by all residual image block r
0, r
1..., r
s-1composition resolution is the residual image R of W × H.Here, the number of behalf residual image block, s=(W × H)/l
2, W represents the width of residual image R, and H represents the height of residual image R;
Step 6, the compressed sensing sampling of residual image
First, to the residual image R produced in step 5, image block method is traditionally divided into N=(W × H)/n
2individual non-overlapping copies, size is the residual image block of n × n, is designated as C
0, C
1..., C
n-1.Here, W represents the width of residual image, and H represents the height of residual image; N represents the number of residual image block; N represents width and the height of residual image block, and n=m, m represent width and the height of the original picture block produced in step 1;
Then, to each residual image block C
0, C
1..., C
n-1realize traditional compressed sensing based on image block one by one with stochastic sampling matrix Ф to sample, obtain sampled data P
0, P
1..., P
n-1; Pooled sampling data P
0, P
1..., P
n-1, be designated as P, P={P
0, P
1..., P
n-1, P is the compressed sensing sampled data of residual image R.
Step 7, the correction of original image compression sampling data
The compression sampling data G of the original image obtained in step 2 is added with the compression sampling data P of the residual image R obtained in step 6, obtains the compression sampling data of revised original image, be designated as Y, here, Y=G+P, wherein Y={Y
0, Y
1..., Y
n-1.Y
0, Y
1..., Y
n-1, be the compression sampling data of revised each image block, here, Y
0=G
0+ P
0, Y
1=G
1+ P
1..., Y
n-1=G
n-1+ P
n-1, the number of N representative image block.
Step 8, the generation of index matrix
First, n
2individual integer 0,1 ..., n
2-1 puts by column by order from small to large, and producing a size is the index matrix of n × n, is designated as A:
Element in A is designated as A (x, y) (0≤x<n, 0≤y<n), and here, n=m, m represent width and the height of the original picture block produced in step 1; X represents the horizontal ordinate of index matrix A interior element, and y represents the ordinate of index matrix A interior element;
Then, with A (0,0) for starting point, the method for matrix being carried out to submatrix division traditionally, is divided into k=n A
2/ l
2individual non-overlapping copies, size is the subindex matrix of l × l, is designated as A
0, A
1..., A
k-1, A also can be expressed as:
By submatrix A
i(i=0,1 ..., k-1) in each element be designated as A
i(x, y) (0≤x<l, 0≤y<l); Here, k represents the number of subindex matrix, k=n
2/ l
2, n=m, m represent width and the height of the original picture block produced in step 1, and l represents line number and the columns of subindex matrix, and l=L, L represent the image that sets in step 4 at equal intervals without the sampling interval distance in the down-sampling of weighting spatial domain; I represents the index subscript of subindex matrix; X represents subindex matrix A
ithe horizontal ordinate of interior element, y represents subindex matrix A
ithe ordinate of interior element;
Step 9, the correction of sampling matrix
First, each column vector of stochastic sampling matrix Ф is expressed as
here, j represents the index subscript of each column vector, j=0, and 1 ..., M, M=m
2, m represents width and the height of the original picture block produced in step 1;
Secondly, the column vector of the M in Ф is divided into k=n
2/ l
2individual Column vector groups, is designated as g
i, here, i represents the index subscript of each Column vector groups, i=0, and 1 ..., k-1, each Column vector groups g
icomprise l
2individual vector, step is as follows:
1st group,
2nd group,
In like manner, to i-th group,
Then, by each Column vector groups g
iinterior all column vectors are added and obtain a column vector, are designated as v
i, step is as follows:
1st group,
2nd group,
In like manner, to i-th group,
here, i represents the index subscript of subindex matrix, i=0, and 1 ..., k-1, k=n
2/ l
2, n=m, m represent width and the height of the original picture block produced in step 1, and l represents line number and the columns of subindex matrix, and l=L, L represent the image that sets in step 4 at equal intervals without the sampling interval distance in the down-sampling of weighting spatial domain;
Finally, k the column vector v will produced
0, v
1..., v
k-1form a sampling matrix, be designated as D, D=[v
0, v
1..., v
k-1].Here, k=n
2/ l
2, represent the number of subindex matrix; N represents line number and the columns of index matrix in step 8, and n=m, m represent width and the height of the original picture block produced in step 1; L represents line number and the columns of subindex matrix, and l=L, L represent the image that sets in step 4 at equal intervals without the sampling interval distance in the down-sampling of weighting spatial domain.
Step 10, the reconstruct of low-resolution image
In the sparse reconstruction method of gradient projection, with sampling matrix D to each the sampled data Y produced in step 7
0, Y
1..., Y
n-1realize traditional image block based on compressed sensing one by one to rebuild, obtain rebuilding the image block that rear size is (n/l) × (n/l), be designated as
by the method for traditional image block composograph by all
form complete reconstruction image, be designated as Q, the resolution of Q is w × h, here, the width of image Q is rebuild in w representative, and the height of image Q is rebuild in h representative, w=W/l, h=H/l, W represent the width of original image, and H represents the height of original image, l=L, L represents the image that sets in step 4 at equal intervals without the sampling interval distance in the down-sampling of weighting spatial domain, and the number of image block is rebuild in N representative, N=(W × H)/n
2, n=m, m represent width and the height of the original picture block produced in step 1.
Innovative point of the present invention: for the existing quality that can reduce the image finally produced based on utilize " reconstruct-spatial domain, compressed sensing territory down-sampling " two steps to realize in the method for compressing image of compressed sensing method that low-resolution image shows, the invention provides a kind of method that new compression directly realizing image in compressed sensing territory and low resolution reconstruct.The method proposed, by revising the compression sampling data of original image, concentrates on compressed sensing territory the Image Reconstruction of low resolution and completes, thus reaches the object improving reconstructed image quality.
Ultimate principle of the present invention: suppose will realize original signal (with discrete signal x
1Dfor example) spatial domain half sample reconstruction, from mathematical meaning, compression sampling reconstruction signal z at this moment
1D=[z
0..., z
l-1]
t∈ R
l, (l=L/2) is to original signal x
1Din finish the part ↓ x after down-sampling
1D=[x
0, x
2, x
l-2]
t∈ R
l/2one approach.Suppose to obtain K (K < < L) individual sample in the sampling process of original signal.Although sampling rate K/L is relative to x
1Dlower, but as l=L/2, sampling rate is relative to ↓ x
1Dthen turn over some, because relative sample rate is now K/l=2 × (K/L).And l is less, relative sample rate K/l will be larger.For 2D signal ↓ x
2D, downward half sampling means and all needs to do half sampling in both direction (vertical and horizontal direction), is so just equivalent to do the sampling of 1/4 to whole signal; And when both direction all does 1/4 sampling, be then equivalent to do the sampling of 1/16 to whole signal.Therefore, even if use very low sampling rate for original image, such as, only have 10%, but for the image done after the down-sampling of spatial domain, sampling rate will double.Such as, for the image after 2 × 2 down-samplings, sampling rate is 40%, and for the image after 4 × 4 down-samplings, sampling rate is 160%, now becomes the situation of over-sampling completely.Exactly because there has been the relative change of this sampling rate, namely high sampling rate sampling or even over-sampling are transformed into from low sampling rate sampling, just for the high-quality low-resolution image of reconstruct provides guarantee.
Technical scheme of the present invention be to compression sampling after sample data revise, and utilize compressed sensing restructing algorithm, reconstruction method as sparse in gradient projection, directly realizing can downscaled images reconstruct.
Particularly, according to compressive sensing theory, to a highly sparse one-dimensional signal x
1D=[x
0..., x
l-1]
t∈ R
lwith one group of (K) linear base, stochastic sampling is carried out to it:
Even if number of samples K < < is L, still can, by solving the problem of a convex optimization, signal be reconstructed out accurately.
For one-dimensional signal, according to formula (1), the sampled data obtained under compressed sensing can be expressed as linear combination:
If supposed
so formula (2) can be write as:
Although each y here
iall from original signal x
1D=[x
0..., x
l-1]
t∈ R
lmiddle acquisition, but it also can be equivalent to from the signal ↓ x after more than half sampling simultaneously
1D=[x
0, x
2, x
l-2]
t∈ R
l/2in directly obtain.Coefficient only now in sampling matrix becomes c
i, 2j+ c
i, 2j+1.Further consider, present original signal is by x
1Dbecome ↓ x
1D.And for ↓ x
1D, sampling rate is improved, therefore, it is possible to rebuild a high-quality signal z
1D.
For 2D signal, the data after sampling can be write as:
Herein
represent dot product, ∑ is to all elements summation in matrix.In order to reconstruct the low-resolution image through 2 × 2 down-samplings, x (2u, 2v) ≡ x (2u+1 can be supposed, 2v) ≡ x (2u, 2v+1) ≡ x (2u+1,2v+1), so formula (4) just can be rewritten as:
Similar with the situation of one dimension, for the situation of two dimension, each sample y
ican be regarded as directly the signal ↓ x after 2 × 2 down-samplings
2Dcarry out the result of stochastic sampling.Each coefficient now in sampling matrix becomes matrix of coefficients [C
m,n]
2 × 2middle all elements and.Similarly, for ↓ x
2D, because sampling rate is improved, a high-quality signal z also can be rebuild
2D=[z (m, n)]
h × w.
No matter for the hypothesis x in one-dimensional problem
2j≡ x
2j+1or for hypothesis x (2u, 2v) ≡ x (2u+1,2v) ≡ x (2u, 2v+1) ≡ x (2u+1,2v+1) in two-dimensional problems, in fact all also one set up surely.If such hypothesis cannot be set up, the data { y so after compression sampling
i↓ x also just cannot be described exactly
1Dor ↓ x
2D.In some extreme cases, describe signal in such a way and even can bring serious error.In order to eliminate this error, need the sample after to sampling to revise, making it approach those can from ↓ x
2Din the sample data (supposing that these data exist) that obtains, namely by amendment { y
iallow numerical value in its regression formulas (5).
Advantage of the present invention; By compressed sensing sampling action after original image, directly realize the low resolution reconstruct of image in compressed sensing territory, avoid the two step implementation procedures by " reconstruct-spatial domain, compressed sensing territory down-sampling ".Essence of the present invention is by first revising the compression sampling data of original image in compressed sensing territory, then is reconstructed, and finally directly produces the image of low resolution in compressed sensing territory.
Accompanying drawing explanation
Fig. 1 is realization flow of the present invention;
Fig. 2 reduces for adopting distinct methods to carry out spatial domain 2 × 2 to different images under different sampling rate PSNR and the SSIM value reconstructing and obtain.
Embodiment
The present invention mainly adopts the mode of emulation experiment to verify the feasibility of this system model, institute in steps, conclusion all on MATLAB7.11 checking correct.Reduce reconstruct for realizing spatial domain 2 × 2, concrete implementation step is as follows:
Step 1, the pre-service of original image
Setting m=32, be the original image of W × H by resolution, image block method is traditionally divided into N=(W × H)/32
2individual non-overlapping copies, size is the image block of 32 × 32, is designated as B
0, B
1..., B
n-1, W represents the width of original image here, and H represents the height of original image, and N represents the number of original picture block.
Step 2, the compressed sensing sampling of image
To each image block B produced in step 1
0, B
1..., B
n-1realize traditional compressed sensing based on image block one by one with stochastic sampling matrix Ф to sample, obtain sampled data G
0, G
1..., G
n-1.Pooled sampling data G
0, G
1..., G
n-1, be designated as G, G={G
0, G
1..., G
n-1, G is the compressed sensing sampled data of original image.Here, the number of N representative image block, the columns of stochastic sampling matrix Ф is 1024, and line number is determined by sampling rate.
Step 3, waits the reconstruct of image in different resolution
Adopt traditional sparse reconstruction method of gradient projection to each the sampled data G obtained in step 2
0, G
1..., G
n-1realize traditional image block based on compressed sensing one by one to rebuild, obtain rebuilding the image block that rear size is 32 × 32, be designated as
will by the method for traditional image block composograph
form complete reconstruction image, be designated as I, the resolution of rebuilding image I is W × H, and here, the width of image I is rebuild in W representative, and the height of image I is rebuild in H representative; The number of image block is rebuild in N representative.
Step 4, carries out piecemeal to reconstruction image
First, to the reconstruction image I produced in step 3, image traditionally, at equal intervals without weighting spatial domain down-sampling method, sets sampling interval distance L=2.Then, reconstruction image I image block method is traditionally divided into s=(W × H)/2
2individual non-overlapping copies, size is the image block of 2 × 2, is designated as b
0, b
1..., b
s-1, and by image block b
iin pixel be designated as b
i(x, y); Here, the width of image I is rebuild in W representative, and the height of image I is rebuild in H representative; The index subscript of image block is rebuild in i representative, i=0, and 1 ..., s-1; The horizontal ordinate of pixel in image block is rebuild in x representative, and the ordinate of pixel in image block is rebuild in y representative, 0≤x<2,0≤y<2.
Step 5, the generation of residual image
First, defining a resolution is the residual image of W × H, is designated as R, and the pixel value of all pixels in R is set as 0; Residual image R image block method is traditionally divided into s=(W × H)/2
2individual non-overlapping copies, size is the image block of 2 × 2, is designated as r
0, r
1..., r
s-1, by image block r
i(i=0,1 ..., s-1) in pixel be designated as r
i(x, y) (0≤x<2,0≤y<2); Here, W represents the width of residual image R, and H represents the height of residual image R; I represents the index subscript of residual image block; X represents the horizontal ordinate of pixel in residual image block, and y represents the ordinate of pixel in residual image block;
Secondly, image block b is rebuild with s that produces in step 4
i(i=0,1 ..., s-1), remove the residual image block r that s correspondence is set
i(i=0,1 ..., s-1) in the pixel value of pixel, step is as follows:
To the 1st residual image block r
0, make r
0(0,0)=b
0(0,0)-b
0(0,0), r
0(0,1)=b
0(0,0)-b
0(0,1), r
0(1,0)=b
0(0,0)-b
0(1,0), r
0(1,1)=b
0(0,0)-b
0(1,1); To the 2nd residual image block r
1, make r
1(0,0)=b
1(0,0)-b
1(0,0), r
1(0,1)=b
1(0,0)-b
1(0,1), r
1(1,0)=b
1(0,0)-b
1(1,0), r
1(1,1)=b
1(0,0)-b
1(1,1); In like manner, to i residual image block, r is made
i-1(0,0)=b
i-1(0,0)-b
i-1(0,0), r
i-1(0,1)=b
i-1(0,0)-b
i-1(0,1), r
i-1(1,0)=b
i-1(0,0)-b
i-1(1,0), r
i-1(1,1)=b
i-1(0,0)-b
i-1(1,1); Here, the index subscript of i representative image block;
Finally, by the method for traditional image block composograph by all residual image block r
0, r
1..., r
s-1composition resolution is the residual image R of W × H.Here, the number of behalf residual image block, s=(W × H)/2
2, W represents the width of residual image R, and H represents the height of residual image R;
Step 6, the compressed sensing sampling of residual image
First, to the residual image R produced in step 5, image block method is traditionally divided into N=(W × H)/32
2individual non-overlapping copies, size is the residual image block of 32 × 32, is designated as C
0, C
1..., C
n-1.Here, W represents the width of residual image, and H represents the height of residual image; N represents the number of residual image block;
Then, to each residual image block C
0, C
1..., C
n-1realize traditional compressed sensing based on image block one by one with stochastic sampling matrix Ф to sample, obtain sampled data P
0, P
1..., P
n-1; Pooled sampling data P
0, P
1..., P
n-1, be designated as P, P={P
0, P
1..., P
n-1, P is the compressed sensing sampled data of residual image R.
Step 7, the correction of original image compression sampling data
The compression sampling data G of the original image obtained in step 2 is added with the compression sampling data P of the residual image R obtained in step 6, obtains the compression sampling data of revised original image, be designated as Y, here, Y=G+P, wherein Y={Y
0, Y
1..., Y
n-1.Y
0, Y
1..., Y
n-1, be the compression sampling data of revised each image block, here, Y
0=G
0+ P
0, Y
1=G
1+ P
1..., Y
n-1=G
n-1+ P
n-1, the number of N representative image block.
Step 8, the generation of index matrix
First, 32
2individual integer 0,1 ..., 1023 put by column by order from small to large, and producing a size is the index matrix of 32 × 32, is designated as A:
Element in A is designated as A (x, y) (0≤x<32,0≤y<32), and here, x represents the horizontal ordinate of index matrix A interior element, and y represents the ordinate of index matrix interior element;
Then, with A (0,0) for starting point, the method for matrix being carried out to submatrix division traditionally, is divided into 32 A
2/ 2
2=256 non-overlapping copies, size is the subindex matrix of 2 × 2, is designated as A
0, A
1..., A
255, A also can be expressed as:
By submatrix A
i(i=0,1 ..., 255) in each element be designated as A
i(x, y) (0≤x<2,0≤y<2); Here, i represents the index subscript of subindex matrix; X represents the horizontal ordinate of subindex matrix interior element, and y represents the ordinate of subindex matrix interior element;
Step 9, the correction of sampling matrix
First, each column vector of stochastic sampling matrix Ф is expressed as
Secondly, the column vector of 1024 in Ф is divided into 256 Column vector groups, is designated as g
i(i=0,1 ..., 255), each Column vector groups g
icomprise 4 vectors, step is as follows:
1st group,
2nd group,
In like manner, to i-th group,
Then, by each Column vector groups g
iinterior all column vectors are added and obtain a column vector, are designated as v
i, step is as follows:
1st group,
2nd group,
In like manner, to i-th group,
here, i represents the index subscript of subindex matrix, i=0, and 1 ..., 255;
Finally, 256 the column vector v will produced
0, v
1..., v
255form a sampling matrix, be designated as D, D=[v
0, v
1..., v
255].
Step 10, the reconstruct of low-resolution image
In the sparse reconstruction method of gradient projection, with sampling matrix D to each the sampled data Y produced in step 7
0, Y
1..., Y
n-1realize traditional image block based on compressed sensing one by one to rebuild, obtain rebuilding the image block that rear size is 16 × 16, be designated as
by the method for traditional image block composograph by all
form complete reconstruction image, be designated as Q, the resolution of Q is w × h, and here, the width of image Q is rebuild in w representative, and the height of image Q is rebuild in h representative, and w=W/2, h=H/2, W represent the width of original image, and H represents the height of original image.
Embodiment is applied to Lena, Fishingboat and Goldhill tri-width is of a size of in the classical legend of 512 × 512, accompanying drawing 2 is under different sampling rate, the contrast that spatial domain 2 × 2 reduces reconstruct is differently carried out to different images, comprise Y-PSNR (peaksignaltonoiseratio, and the comparison of structural similarity (structuralsimilarity, SSIM) PSNR).Clearly, the method in the present invention has obvious performance boost than existing methods.
Claims (1)
1. the spatial domain based on compressed sensing sampling data correction can downscaled images reconstructing method, it is characterized in that it comprises the following steps:
Step 1, the pre-service of original image
Be the original image of W × H by resolution, image block method is traditionally divided into N=(W × H)/m
2individual non-overlapping copies, size is the image block of m × m, is designated as B
0, B
1..., B
n-1, W represents the width of original image here, and H represents the height of original image, and N represents the number of original picture block, and m represents width and the height of original picture block;
Step 2, the compressed sensing sampling of image
To each image block B produced in step 1
0, B
1..., B
n-1realize traditional compressed sensing based on image block one by one with stochastic sampling matrix Ф to sample, obtain sampled data G
0, G
1..., G
n-1; Pooled sampling data G
0, G
1..., G
n-1, be designated as G, G={G
0, G
1..., G
n-1, G is the compressed sensing sampled data of original image; Here, the number of N representative image block, the columns of stochastic sampling matrix Ф is m
2, line number is determined by the sampling rate in the sampling of traditional compressed sensing;
Step 3, waits the reconstruct of image in different resolution
Adopt traditional sparse reconstruction method of gradient projection to each the sampled data G obtained in step 2
0, G
1..., G
n-1realize traditional image block based on compressed sensing one by one to rebuild, obtain rebuilding the image block that rear size is n × n, be designated as
will by the method for traditional image block composograph
form complete reconstruction image, be designated as I, the resolution of rebuilding image I is W × H, and here, the width of image I is rebuild in W representative, and the height of image I is rebuild in H representative; The number of image block is rebuild in N representative; Width and the height of image block are rebuild in n representative, and n=m, m represent width and the height of the original picture block produced in step 1;
Step 4, carries out piecemeal to reconstruction image
First, to the reconstruction image I produced in step 3, image is traditionally at equal intervals without weighting spatial domain down-sampling method, and setting sampling interval distance, is designated as L, L=2
d<m, d are positive integer, and m represents width and the height of the original picture block produced in step 1; Then, reconstruction image I image block method is traditionally divided into s=(W × H)/l
2individual non-overlapping copies, size is the image block of l × l, is designated as b
0, b
1..., b
s-1, and by image block b
iin pixel be designated as b
i(x, y); Here, the width of image I is rebuild in W representative, and the height of image I is rebuild in H representative; Width and the height of image block are rebuild in l representative, l=L; The index subscript of image block is rebuild in i representative, i=0, and 1 ..., s-1; The horizontal ordinate of pixel in image block is rebuild in x representative, and the ordinate of pixel in image block is rebuild in y representative, 0≤x<l, 0≤y<l;
Step 5, the generation of residual image
First, defining a resolution is the residual image of W × H, is designated as R, and the pixel value of all pixels in R is set as 0; Residual image R image block method is traditionally divided into s=(W × H)/l
2individual non-overlapping copies, size is the image block of l × l, is designated as r
0, r
1..., r
s-1, by image block r
iin pixel be designated as r
i(x, y); Here, i=0,1 ..., s-1,0≤x<l, 0≤y<l, W represent the width of residual image R, and H represents the height of residual image R; L represents width and the height of residual image block, and l=L, L represent the image that sets in step 4 at equal intervals without the sampling interval distance in the down-sampling of weighting spatial domain; I represents the index subscript of residual image block; X represents the horizontal ordinate of pixel in residual image block, and y represents the ordinate of pixel in residual image block;
Secondly, image block b is rebuild with s that produces in step 4
i, i=0,1 ..., s-1, arranges the residual image block r of s correspondence
iin the pixel value of pixel, i=0,1 ..., s-1, step is as follows:
To the 1st residual image block r
0, make r
0(0,0)=b
0(0,0)-b
0(0,0), r
0(0,1)=b
0(0,0)-b
0(0,1) ..., r
0(l-1, l-1)=b
0(0,0)-b
0(l-1, l-1); To the 2nd residual image block r
1, make r
1(0,0)=b
1(0,0)-b
1(0,0), r
1(0,1)=b
1(0,0)-b
1(0,1) ..., r
1(l-1, l-1)=b
1(0,0)-b
1(l-1, l-1); In like manner, to i residual image block, r is made
i-1(0,0)=b
i-1(0,0)-b
i-1(0,0), r
i-1(0,1)=b
i-1(0,0)-b
i-1(0,1) ..., r
i-1(l-1, l-1)=b
i-1(0,0)-b
i-1(l-1, l-1); Here, the index subscript of i representative image block; The width of l representative image block and height, l=L, L represent the image that sets in step 4 at equal intervals without the sampling interval distance in the down-sampling of weighting spatial domain; The number of behalf image block, s=(W × H)/l
2, W represents the width of residual image R, and H represents the height of residual image R;
Finally, by the method for traditional image block composograph by all residual image block r
0, r
1..., r
s-1composition resolution is the residual image R of W × H; Here, the number of behalf residual image block, s=(W × H)/l
2, W represents the width of residual image R, and H represents the height of residual image R;
Step 6, the compressed sensing sampling of residual image
First, to the residual image R produced in step 5, image block method is traditionally divided into N=(W × H)/n
2individual non-overlapping copies, size is the residual image block of n × n, is designated as C
0, C
1..., C
n-1; Here, W represents the width of residual image, and H represents the height of residual image; N represents the number of residual image block; N represents width and the height of residual image block, and n=m, m represent width and the height of the original picture block produced in step 1;
Then, to each residual image block C
0, C
1..., C
n-1realize traditional compressed sensing based on image block one by one with stochastic sampling matrix Ф to sample, obtain sampled data P
0, P
1..., P
n-1; Pooled sampling data P
0, P
1..., P
n-1, be designated as P, P={P
0, P
1..., P
n-1, P is the compressed sensing sampled data of residual image R;
Step 7, the correction of original image compression sampling data
The compression sampling data G of the original image obtained in step 2 is added with the compression sampling data P of the residual image R obtained in step 6, obtains the compression sampling data of revised original image, be designated as Y, here, Y=G+P, wherein Y={Y
0, Y
1..., Y
n-1; Y
0, Y
1..., Y
n-1, be the compression sampling data of revised each image block, here, Y
0=G
0+ P
0, Y
1=G
1+ P
1..., Y
n-1=G
n-1+ P
n-1, the number of N representative image block;
Step 8, the generation of index matrix
First, n
2individual integer 0,1 ..., n
2-1 puts by column by order from small to large, and producing a size is the index matrix of n × n, is designated as A, and n is positive integer,
Element in A is designated as A (x, y), here, 0≤x<n, 0≤y<n, n=m, m represent width and the height of the original picture block produced in step 1; X represents the horizontal ordinate of index matrix A interior element, and y represents the ordinate of index matrix A interior element;
Then, with A (0,0) for starting point, the method for matrix being carried out to submatrix division traditionally, is divided into k=n A
2/ l
2individual non-overlapping copies, size is the subindex matrix of l × l, is designated as A
0, A
1..., A
k-1, A also can be expressed as:
By submatrix A
iin each element be designated as A
i(x, y); Here, i=0,1 ..., k-1,0≤x<l, 0≤y<l, k represent the number of subindex matrix, k=n
2/ l
2, n=m, m represent width and the height of the original picture block produced in step 1, and l represents line number and the columns of subindex matrix, and l=L, L represent the image that sets in step 4 at equal intervals without the sampling interval distance in the down-sampling of weighting spatial domain; I represents the index subscript of subindex matrix; X represents subindex matrix A
ithe horizontal ordinate of interior element, y represents subindex matrix A
ithe ordinate of interior element;
Step 9, the correction of sampling matrix
First, each column vector of stochastic sampling matrix Ф is expressed as
here, j represents the index subscript of each column vector, j=0, and 1 ..., M, M=m
2, m represents width and the height of the original picture block produced in step 1;
Secondly, the column vector of the M in Ф is divided into k=n
2/ l
2individual Column vector groups, is designated as g
i, here, i represents the index subscript of each Column vector groups, i=0, and 1 ..., k-1, each Column vector groups g
icomprise l
2individual vector, step is as follows:
1st group,
2nd group,
In like manner, to i-th group,
Then, by each Column vector groups g
iinterior all column vectors are added and obtain a column vector, are designated as v
i, step is as follows:
1st group,
2nd group,
In like manner, to i-th group,
here, i represents the index subscript of subindex matrix, i=0, and 1 ..., k-1, k=n
2/ l
2, n=m, m represent width and the height of the original picture block produced in step 1, and l represents line number and the columns of subindex matrix, and l=L, L represent the image that sets in step 4 at equal intervals without the sampling interval distance in the down-sampling of weighting spatial domain;
Finally, k the column vector v will produced
0, v
1..., v
k-1form a sampling matrix, be designated as D, D=[v
0, v
1..., v
k-1]; Here, k=n
2/ l
2, represent the number of subindex matrix; N represents line number and the columns of index matrix in step 8, and n=m, m represent width and the height of the original picture block produced in step 1; L represents line number and the columns of subindex matrix, and l=L, L represent the image that sets in step 4 at equal intervals without the sampling interval distance in the down-sampling of weighting spatial domain;
Step 10, the reconstruct of low-resolution image
In the sparse reconstruction method of gradient projection, with sampling matrix D to each the sampled data Y produced in step 7
0, Y
1..., Y
n-1realize traditional image block based on compressed sensing one by one to rebuild, obtain rebuilding the image block that rear size is (n/l) × (n/l), be designated as
by the method for traditional image block composograph by all
form complete reconstruction image, be designated as Q, the resolution of Q is w × h, here, the width of image Q is rebuild in w representative, and the height of image Q is rebuild in h representative, w=W/l, h=H/l, W represent the width of original image, and H represents the height of original image, l=L, L represents the image that sets in step 4 at equal intervals without the sampling interval distance in the down-sampling of weighting spatial domain, and the number of image block is rebuild in N representative, N=(W × H)/n
2, n=m, m represent width and the height of the original picture block produced in step 1.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310421657.0A CN103473797B (en) | 2013-09-16 | 2013-09-16 | Spatial domain based on compressed sensing sampling data correction can downscaled images reconstructing method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310421657.0A CN103473797B (en) | 2013-09-16 | 2013-09-16 | Spatial domain based on compressed sensing sampling data correction can downscaled images reconstructing method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103473797A CN103473797A (en) | 2013-12-25 |
| CN103473797B true CN103473797B (en) | 2016-04-20 |
Family
ID=49798630
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310421657.0A Expired - Fee Related CN103473797B (en) | 2013-09-16 | 2013-09-16 | Spatial domain based on compressed sensing sampling data correction can downscaled images reconstructing method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103473797B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP7002056B2 (en) * | 2016-09-12 | 2022-02-10 | パナソニックＩｐマネジメント株式会社 | 3D model generator and 3D model generation method |
| CN106530253B (en) * | 2016-11-09 | 2019-05-03 | 西南科技大学 | A kind of construction method of SAR image compressed sensing reconstruction sample |
| CN107146259B (en) * | 2017-04-05 | 2019-12-10 | 电子科技大学 | Novel image compression method based on compressed sensing theory |
| CN109101504B (en) * | 2017-06-20 | 2023-09-19 | 恒为科技（上海）股份有限公司 | Log compression and indexing method |
| CN108010000B (en) * | 2017-12-12 | 2021-11-02 | 南京邮电大学 | Compressed image sharpening method based on adjacent feature subspace fitting |
| CN108447101B (en) * | 2018-01-23 | 2019-08-13 | 电子科技大学 | A kind of compression of images cognitive method based on JND model |
| CN115662346B (en) * | 2022-10-14 | 2024-01-19 | 格兰菲智能科技有限公司 | Demura compensation value compression method and system |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101895297A (en) * | 2010-07-30 | 2010-11-24 | 哈尔滨工业大学 | Compressed sensing-oriented block-sparse signal reconfiguring method |
| CN102142133A (en) * | 2011-04-19 | 2011-08-03 | 西安电子科技大学 | Mammary X-ray image enhancement method based on non-subsampled Directionlet transform and compressive sensing |
| CN102629373A (en) * | 2012-02-27 | 2012-08-08 | 天津大学 | Super-resolution image acquisition method based on sparse representation theory |
| CN102750677A (en) * | 2012-06-12 | 2012-10-24 | 清华大学 | Single image super-resolution method based on identical scale structure self-similarity and compressed sensing |
| CN103077510A (en) * | 2013-01-21 | 2013-05-01 | 中国计量学院 | Multivariate compressive sensing reconstruction method based on wavelet HMT (Hidden Markov Tree) model |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8897515B2 (en) * | 2009-09-08 | 2014-11-25 | Mayo Foundation For Medical Education And Research | Method for compressed sensing image reconstruction using a priori knowledge of spatial support |
| CN102959388B (en) * | 2010-06-24 | 2016-02-24 | 皇家飞利浦电子股份有限公司 | Utilize the dynamic contrast Enhanced MR imaging that compression sensing is rebuild |
| US8570442B2 (en) * | 2011-07-12 | 2013-10-29 | Xerox Corporation | Hyperspectral image reconstruction via a compressed sensing framework |
-
2013
- 2013-09-16 CN CN201310421657.0A patent/CN103473797B/en not_active Expired - Fee Related
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101895297A (en) * | 2010-07-30 | 2010-11-24 | 哈尔滨工业大学 | Compressed sensing-oriented block-sparse signal reconfiguring method |
| CN102142133A (en) * | 2011-04-19 | 2011-08-03 | 西安电子科技大学 | Mammary X-ray image enhancement method based on non-subsampled Directionlet transform and compressive sensing |
| CN102629373A (en) * | 2012-02-27 | 2012-08-08 | 天津大学 | Super-resolution image acquisition method based on sparse representation theory |
| CN102750677A (en) * | 2012-06-12 | 2012-10-24 | 清华大学 | Single image super-resolution method based on identical scale structure self-similarity and compressed sensing |
| CN103077510A (en) * | 2013-01-21 | 2013-05-01 | 中国计量学院 | Multivariate compressive sensing reconstruction method based on wavelet HMT (Hidden Markov Tree) model |
Non-Patent Citations (5)
| Title |
|---|
| Compressed sensing technique for high-resolution radar imaging;Yeo-Sun Yoon 等;《Signal Processing, Sensor Fusion, and Target Recognition》;20080417;第69681A.1-69681A.10页 * |
| k-t FOCUSS: A General Compressed Sensing Framework;Hong Jung 等;《Magnetic Resonance in Medicine》;20090131;第61卷(第1期);第103-116页 * |
| Residual Reconstruction for Block-Based Compressed Sensing of Video;Sungkwang Mun 等;《Data Compression Conference》;20110331;第29-31页 * |
| 基于压缩感知与结构自相似性的遥感图像超分辨率方法;潘宗序 等;《信号处理》;20120630;第28卷(第6期);第859-872页 * |
| 基于词典学习和稀疏表示的超分辨率方法;浦剑 等;《模式识别与人工智能》;20100131;第23卷(第3期);第335-340页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103473797A (en) | 2013-12-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103473797B (en) | Spatial domain based on compressed sensing sampling data correction can downscaled images reconstructing method | |
| CN103473744B (en) | Spatial domain based on the sampling of variable weight formula compressed sensing can downscaled images reconstructing method | |
| Zhu et al. | Fast single image super-resolution via self-example learning and sparse representation | |
| CN110969577B (en) | Video super-resolution reconstruction method based on deep double attention network | |
| CN111260586B (en) | Correction method and device for distorted document image | |
| CN112700392A (en) | Video super-resolution processing method, device and storage medium | |
| CN103020909B (en) | Single-image super-resolution method based on multi-scale structural self-similarity and compressive sensing | |
| CN105046672A (en) | Method for image super-resolution reconstruction | |
| CN105741252A (en) | Sparse representation and dictionary learning-based video image layered reconstruction method | |
| CN105631807A (en) | Single-frame image super resolution reconstruction method based on sparse domain selection | |
| Wei et al. | Improving resolution of medical images with deep dense convolutional neural network | |
| CN111028153A (en) | Image processing and neural network training method and device and computer equipment | |
| CN106327422A (en) | Image stylized reconstruction method and device | |
| Ma et al. | Sparse representation and position prior based face hallucination upon classified over-complete dictionaries | |
| CN112529776A (en) | Training method of image processing model, image processing method and device | |
| CN104899835A (en) | Super-resolution processing method for image based on blind fuzzy estimation and anchoring space mapping | |
| Wu et al. | Cross-view panorama image synthesis with progressive attention GANs | |
| CN115272082A (en) | Model training method, video quality improving method, device and computer equipment | |
| CN103390266B (en) | A kind of image super-resolution method and device | |
| CN102842123B (en) | Sparse-region residual error compensating and revising method for improving marginal definition during image sampling | |
| CN102930573B (en) | A kind of image rebuilding method based on the sparse model of two-dimension analysis and training dictionary thereof | |
| CN104091364A (en) | Single-image super-resolution reconstruction method | |
| CN102708556A (en) | Single image super resolution method on basis of maintenance of reversed image | |
| Gao et al. | Multi-branch aware module with channel shuffle pixel-wise attention for lightweight image super-resolution | |
| CN116152710A (en) | Video instance segmentation method based on cross-frame instance association |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20160420 Termination date: 20190916 |