CN106204482A - Based on the mixed noise minimizing technology that weighting is sparse - Google Patents
Based on the mixed noise minimizing technology that weighting is sparse Download PDFInfo
- Publication number
- CN106204482A CN106204482A CN201610538475.5A CN201610538475A CN106204482A CN 106204482 A CN106204482 A CN 106204482A CN 201610538475 A CN201610538475 A CN 201610538475A CN 106204482 A CN106204482 A CN 106204482A
- Authority
- CN
- China
- Prior art keywords
- image
- sparse
- block
- non local
- denoising
- 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
- 238000005516 engineering process Methods 0.000 title claims abstract description 11
- 238000000034 method Methods 0.000 claims abstract description 65
- 206010008190 Cerebrovascular accident Diseases 0.000 claims description 4
- 208000006011 Stroke Diseases 0.000 claims description 4
- 230000009977 dual effect Effects 0.000 claims description 4
- 239000011159 matrix material Substances 0.000 claims description 4
- 230000003044 adaptive effect Effects 0.000 claims description 3
- 230000000694 effects Effects 0.000 abstract description 8
- 230000002401 inhibitory effect Effects 0.000 abstract description 3
- 238000012549 training Methods 0.000 abstract description 3
- 238000012360 testing method Methods 0.000 description 4
- 235000002566 Capsicum Nutrition 0.000 description 3
- 241000758706 Piperaceae Species 0.000 description 3
- 238000001514 detection method Methods 0.000 description 3
- 230000000007 visual effect Effects 0.000 description 3
- 235000013399 edible fruits Nutrition 0.000 description 2
- 238000000513 principal component analysis Methods 0.000 description 2
- 210000000707 wrist Anatomy 0.000 description 2
- 230000006978 adaptation Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000003709 image segmentation Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/28—Determining representative reference patterns, e.g. by averaging or distorting; Generating dictionaries
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Probability & Statistics with Applications (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
The present invention discloses a kind of mixed noise minimizing technology sparse based on weighting, it adds the calculus of variations on the basis of weighting rarefaction representation non local training dictionary, non local similar block is mated, then solves mixing denoising image by the method for antithesis, can preferably preserve the marginal information of image.The denoising effect of the present invention is better than existing algorithm, and it has the highest Y-PSNR and characteristics of image similarity, and mixed noise is had good inhibiting effect, particularly can be good at preserving the marginal information of image, and the reservation to characteristics of image has some improvement.
Description
Technical field
The invention belongs to digital image processing techniques field, be specifically related to a kind of based on weighting sparse mixed noise
Except method.
Background technology
Image is obtaining and is inevitably being polluted by noise during transport, and by the image of sound pollution
The subsequent treatment of image, such as rim detection, target recognition and image segmentation etc. can be affected.Therefore denoising is image digital image
The major issue processed, its main mesh is under preserving the features such as the details of image, texture and edge, from given noise
Image recovers preferable image.Owing to mixed noise does not has parameter model and complex distribution, effective mixed noise of removing is
One extremely difficult problem.
In recent years, people are devoted for years in the improvement removing mixed noise method, as Xiao etc. (Y.Xiao, T.Y.Zeng,
J.Yu and M.K.Ng, " Restoration of images corrupted by mixed Guassian-impulse
noise via l1-l0Minimization, " Pattern Recognit., vol.44, no.8, pp.1708-1720, Aug,
2010) carry out impulse noise detection first with the filtering of intermediate value class, finally use l1-l0Minimum optimization problem denoising image, the party
Though method improves the visual quality of denoising image, but computationally intensive.(J.Jiang, L.Zhang, the and such as Jiang in 2014
J.Yang, " Mixed noise removal by weighted encoding with sparse nonlocal
Regularization, " IEEE Trans.Image Process., vol.23, no.6, pp.2651-2262,2014) (JZY
Model) combine the method that non-local sparse represents, though this algorithm can be removed by white Gaussian noise and impulsive noise and not simultaneously
Need the detection of advanced horizontal pulse noise, but the edge that the method makes denoising figure is fuzzyyer.Simultaneously Zhang etc. (J.Zhang,
D.B.Zhao, and W.Gao, " Group-based sparse representation for image restoration, "
IEEE Trans.Image Process., vol.23, no.8, pp.3336-3351,2014) (ZZG model) propose based on packet
The method of rarefaction representation removes mixed noise, but the denoising figure that the method obtains is the most smooth.
Summary of the invention
The technical problem to be solved is to provide a kind of mixed noise minimizing technology sparse based on weighting, and it will
Non-local sparse weighted coding combines with variation, it is thus possible to preferably preserve the marginal information of image.
For solving the problems referred to above, the present invention is achieved by the following technical solutions:
A kind of mixed noise minimizing technology sparse based on weighting, comprises the following steps that
Step 1, obtains natural image, and the pretreatment being normalized image in standard picture storehouse, then carries out figure
As piecemeal, each image block sparse coding αiRepresent；
Step 2, carries out plus noise process to each image, utilizes the method for variation and weight sparse coding to obtain removal mixed
Close noise model；
Step 3, carries out non local Similarity matching for each image block, finds the similar image block with identical structure type
As the non local similar block of this image block, and obtain the non local factor mu of each image blocki；
Step 4, for each image block, carries out k mean cluster to its non local similar block, non local similar block is clustered
It is divided into k class, and trains k adaptive sparse dictionary D=[D at each apoplexy due to endogenous wind1,D2,…,Dk]；
Step 5, utilizes the method for iteration weight to update the sparse coding α of each image blocki；
After step 6, sparse dictionary and sparse coding determine, utilize the removal mixing that Dual Method solution procedure 2 is constructed
Noise model, obtains denoising image；
Above-mentioned i represents the i-th piecemeal of image, i=1,2 ..., n, n are image block number；K is setting value, for more than or equal to 1
Positive integer.
Compared with prior art, the invention has the beneficial effects as follows:
1, rarefaction representation is trained dictionary to introduce and is removed mixed noise, with minimum whole image of element representation by the present invention
Information, so greatly reduces workload, removes mixed noise by study dictionary, so more effective denoising.
2, non local regularization is introduced and removes mixed noise by the present invention, is conducive to the contact between contact image block, adds
Strong denoising.
3, variation item is introduced and removes mixed noise by the present invention, is conducive to retaining the marginal information of denoising image, improves figure
The visual effect of picture, improves the quality of image.
4, emulation experiment shows, the present invention not only has good inhibiting effect to mixed noise, and to characteristics of image
Reservation has some improvement.
Accompanying drawing explanation
Fig. 1 is the flow chart of a kind of mixed noise minimizing technology sparse based on weighting.
Fig. 2 be the figure of entitled " Pirate " be 30 in white Gaussian noise standard deviation, under the close p=40% of impulsive noise, no
Effect contrast figure with denoising method.Wherein, (a) is original image and partial enlarged drawing；B () is the denoising figure drawn game of ZZG method
Portion's enlarged drawing；C () is denoising figure and the partial enlarged drawing of JZY method, (d) is the denoising figure of the inventive method.
Fig. 3 be the figure of entitled " Peppers " be 20 in white Gaussian noise standard deviation, under impulsive noise density p=50%,
The effect contrast figure of different denoising methods.Wherein, (a) is original image and partial enlarged drawing；(b) be ZZG method denoising figure and
Partial enlarged drawing；C () is denoising figure and the partial enlarged drawing of JZY method, (d) is the denoising figure of the inventive method.
Detailed description of the invention
A kind of mixed noise minimizing technology sparse based on weighting, as it is shown in figure 1, comprise the following steps that
Step 1, obtains natural image, and the pretreatment being normalized image in standard picture storehouse, then carries out figure
As piecemeal.
In standard picture storehouse obtain natural image, the size all 512 × 512 of image, gray value 0-255 it
Between, in order to simplify amount of calculation, every width figure is all normalized, after normalized, again image is carried out piecemeal, every piece
The size present invention be divided into 7 × 7, then every image just obtains 49 fritters.
Step 2, carries out plus noise process to each image, utilizes the method for variation and weight sparse coding to obtain removal mixed
Close noise model.
First, each image is carried out plus noise process, if y is ∈ RnFor noisy observation figure, x ∈ RnFor denoising figure, n is
White Gaussian noise, then the mathematical formulae of white Gaussian noise is that y=x+n. sets the dynamic range of y as [dmin,dmax], impulsive noise
Probability is p (0≤p≤1), then mixed noise model is:
We add white Gaussian noise standard deviation respectively to image is 10,20,30, and impulsive noise density p=30%, 40%,
50%.
Then, at non local weight sparse coding item(JZY) increase on the basis of method
Preserve variation item | | x | | of image borderTVObtain remove mixed noise model:
Wherein, W is weight, and D is sparse dictionary, and α is sparse coding, and μ is the non local factor.
Step 3, carries out non local Similarity matching for each image block, finds it to have the similar image of identical structure type
Block, i.e. its non local similar block.
First image x is carried out piecemeal xiRepresent, each piece of image xiInformation sparse coding αiRepresenting, wherein i represents
I-th piecemeal of image (i=1,2 ..., n), i.e. sparse coding α=[α1,α2,…,αn], so sparse dictionary D=[D1,D2,…,
Dn], i.e. xi=D αi, non local factor μ of corresponding blocksiRepresent, μ=[μ1,μ2,…,μn] so μi=∑ ωi,jαi,j, wherein
αi,jRepresent the similar block finding block j to block i, ωi,jFor weight,xi,j=D αi,jH is for giving
Fixed scalar, it is normalization factor that the present invention arranges h=80, ω, and we use xiWith xi,jBetween Euclidean distance define image
Similarity between block and block, the most how to be defined into the end have how many pieces similar to given image block, be to be defined by oneself,
The present invention is provided with 15 blocks similar to given block.
Step 4, for given each piecemeal, first carries out k mean cluster to the non local similar block of image, exists respectively
Every apoplexy due to endogenous wind updates adaptive sparse PCA (principal component analysis) dictionary.
(4.1) apply in the non local similar block of step 3 by the method for k mean cluster, thus non local similar
Block cluster is k class, is designated as 1,2,3 ..., k, in the present invention k=49.
(4.2) constructing sparse sub-dictionary at each apoplexy due to endogenous wind, thus k the sub-dictionary of self adaptation of training, is designated as D respectively1,
D2..., Dk, k sub-dictionary composition sparse dictionary D=[D1,D2,…,Dk], sparse dictionary has in the vision and quality of image
Well improve.
Step 5: utilize the method for iteration weight to obtain sparse coding.
The method utilizing iteration weight obtains sparse coding, the most first assumes that V is known diagonal matrix function, given V's
Initial value is unit matrix, each diagonal element V of diagonal matrix ViiRepresent, then we set (k '+1) secondary iterationWherein λ is nonnegative constant, and ε=y-x non-negative is error, is obtained by the method for iteration weight
To (k '+1) secondary iteration sparse coding:
α(k′+1)=(DTWD+V(k′+1))-1(DTWy-DTWDμ)+μ。
Step 6: utilize Dual Method solving model, obtains denoising image.
After sparse dictionary D and sparse coding α updates well, the image after utilizing the method for antithesis to obtain denoising, i.e.
Object functionIt is converted into corresponding dual problem i.e.x
× x=X ∈ Rn×n, p ∈ X, and | | p | |∞≤ 1, so we just obtain final removal mixed noise image x.
The effect of the present invention is further illustrated by following emulation.
1. simulated conditions
(1) choose natural image in the standard testing image storehouse that size is 512 × 512 to test；
(2) piece image is divided into the image block of 7 × 7, and the size wherein crossing complete dictionary is 49, and the image of test adds mixed
Close noise be white Gaussian noise standard deviation be 10,20,30, impulsive noise density p=30%, 40%, 50%.
2. emulation content and result
Emulation content: utilize " Pirate " figure and " Peppers " figure of 512 × 512, by the inventive method and prior art
ZZG method, JZY method respectively to its remove mixed noise.
Experimental result:
(1) present invention obtains good denoising effect than ZZG method and JZY method on Y-PSNR PSNR and FSIM
Really.Experimental result is as shown in table 1.As can be seen from the table, the present invention is compared with ZZG method and JZY method, at different noises
With on test image, all achieve higher PSNR value and FSIM value.
The PSNR (dB) and FSIM of table 1 different mixing noise remove method
(2) present invention obtains more preferable denoising effect than ZZG method and JZY method in visual effect.Experimental result is such as
Fig. 2 and as shown in Figure 3.
Fig. 2 white Gaussian noise standard deviation is 30, " Pirate " denoising figure of impulsive noise density p=50%.A () is artwork
Picture and partial enlarged drawing.B () is denoising figure and the partial enlarged drawing of ZZG method, the comparison of the image procossing of ZZG method obscures,
The edge that denoising figure there is also a lot of mixed noise, face and wrist is the most unintelligible.C () is denoising figure and the local of JZY method
Enlarged drawing, although JZY method is better than ZZG method at Y-PSNR and characteristics of image similarity, and the denoising effect of image is also
It is better than ZZG method, but the face mask of ZZG method denoising figure is not it is obvious that the chain in wrist and the ring on finger
Finger is not to will be apparent from.D () is the denoising figure of the inventive method.As can be seen here, the inventive method can be good at preserving image
Marginal information.
Fig. 3 be white Gaussian noise standard deviation be 20, " Peppers " denoising figure of impulsive noise density p=40%.(a) be
Original image and partial enlarged drawing.B () is denoising figure and the partial enlarged drawing of ZZG method, the denoising figure smoother of ZZG method,
The particularly Fructus Capsici base of a fruit.C () is the denoising figure of JZY method, the edge treated of the Fructus Capsici base of a fruit do not have ZZG method good.D () is the present invention
The denoising figure of method.As can be seen here, the inventive method can be good at preserving the characteristic information of image.
The present invention adds the calculus of variations on the basis of weighting rarefaction representation non local training dictionary, enters non local similar block
Row coupling, then solves mixing denoising image by the method for antithesis, can preferably preserve the marginal information of image.The present invention goes
Effect of making an uproar is better than existing algorithm, and it has the highest Y-PSNR and characteristics of image similarity, has mixed noise well
Inhibitory action, particularly can be good at preserving the marginal information of image, and the reservation to characteristics of image has some improvement.
Claims (3)
1., based on the mixed noise minimizing technology that weighting is sparse, it is characterized in that, comprise the following steps that
Step 1, obtains natural image, and the pretreatment being normalized image in standard picture storehouse, then carries out image and divide
Block, each image block sparse coding represents；
Step 2, carries out plus noise process to each image, utilizes the method for variation and weight sparse coding to obtain removing mixing and makes an uproar
Acoustic model；
Step 3, carries out non local Similarity matching for each image block, finds the similar image block conduct with identical structure type
The non local similar block of this image block, and obtain the non local factor of each image block；
Step 4, for each image block, carries out k mean cluster to its non local similar block, and non local similar block cluster is divided into
K class, and train k adaptive sparse dictionary at each apoplexy due to endogenous wind；Wherein k is setting value, for the positive integer more than or equal to 1；
Step 5, utilizes the method for iteration weight to update the sparse coding of each image block；
After step 6, sparse dictionary and sparse coding determine, utilize the removal mixed noise that Dual Method solution procedure 2 is constructed
Model, obtains denoising image.
A kind of mixed noise minimizing technology sparse based on weighting the most according to claim 1, is characterized in that, in step 2,
Removing mixed noise model is:
Wherein, W is the weight set, and y is given noisy image, and D is sparse dictionary, and α is sparse coding, αiFor phase after piecemeal
Answer the sparse coding of block, μiFor the non local factor of relevant block after piecemeal, x is denoising image,It is 2 norms, | | | |1It is 1
Norm, | | x | |TVFull variation for denoising image.
A kind of mixed noise minimizing technology sparse based on weighting the most according to claim 1, is characterized in that, in step 5,
The more new formula of sparse coding is:
α(k′+1)=(DTWD+V(k′+1))-1(DTWy-DTWDμ)+μ
Wherein, α is sparse coding, and D is sparse dictionary, and W is the weight set, and V is given diagonal matrix, and μ is non local dilute
Dredging encoding Factor, y is given noisy image, and k ' is iterations.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610538475.5A CN106204482B (en) | 2016-07-08 | 2016-07-08 | Based on the mixed noise minimizing technology that weighting is sparse |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610538475.5A CN106204482B (en) | 2016-07-08 | 2016-07-08 | Based on the mixed noise minimizing technology that weighting is sparse |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106204482A true CN106204482A (en) | 2016-12-07 |
| CN106204482B CN106204482B (en) | 2019-03-29 |
Family
ID=57474176
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610538475.5A Active CN106204482B (en) | 2016-07-08 | 2016-07-08 | Based on the mixed noise minimizing technology that weighting is sparse |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106204482B (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107085839A (en) * | 2017-06-14 | 2017-08-22 | 西安电子科技大学 | SAR image method for reducing speckle with sparse coding is strengthened based on texture |
| CN107358589A (en) * | 2017-07-17 | 2017-11-17 | 桂林电子科技大学 | A kind of combination histogram of gradients and the denoising method of low-rank constraint |
| CN107730468A (en) * | 2017-10-23 | 2018-02-23 | 鹰艾思科技(深圳)有限公司 | The restoration methods of picture rich in detail under a kind of UAV Fuzzy noise image |
| CN108710150A (en) * | 2018-05-22 | 2018-10-26 | 中国海洋石油集团有限公司 | A kind of earthquake random noise minimizing technology based on steady singular spectrum analysis |
| CN108765322A (en) * | 2018-05-16 | 2018-11-06 | 上饶师范学院 | Image de-noising method and device |
| CN111402173A (en) * | 2020-03-24 | 2020-07-10 | 北京工业大学 | Hybrid noise removing method and device, electronic equipment and storage medium |
| CN111586262A (en) * | 2020-03-27 | 2020-08-25 | 成都航天通信设备有限责任公司 | Image denoising method based on field army image transmission |
| CN113192078A (en) * | 2021-04-19 | 2021-07-30 | 清华大学 | Rapid sea surface SAR image superpixel segmentation method based on density |
| CN113487522A (en) * | 2021-09-08 | 2021-10-08 | 深圳市诚识科技有限公司 | Multi-channel switching noise reduction method for image communication |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102184533A (en) * | 2011-06-10 | 2011-09-14 | 西安电子科技大学 | Non-local-restriction-based total variation image deblurring method |
| US20130156340A1 (en) * | 2011-12-20 | 2013-06-20 | Fatih Porikli | Image Filtering by Sparse Reconstruction on Affinity Net |
| CN103854262A (en) * | 2014-03-20 | 2014-06-11 | 西安电子科技大学 | Medical image noise reduction method based on structure clustering and sparse dictionary learning |
| CN104063886A (en) * | 2014-03-24 | 2014-09-24 | 杭州电子科技大学 | Nuclear magnetic resonance image reconstruction method based on sparse representation and non-local similarity |
| CN104657951A (en) * | 2015-03-02 | 2015-05-27 | 桂林电子科技大学 | Multiplicative noise removal method for image |
| CN105513026A (en) * | 2015-12-17 | 2016-04-20 | 浙江大学 | Compressed sensing reconstruction method based on image nonlocal similarity |
-
2016
- 2016-07-08 CN CN201610538475.5A patent/CN106204482B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102184533A (en) * | 2011-06-10 | 2011-09-14 | 西安电子科技大学 | Non-local-restriction-based total variation image deblurring method |
| US20130156340A1 (en) * | 2011-12-20 | 2013-06-20 | Fatih Porikli | Image Filtering by Sparse Reconstruction on Affinity Net |
| CN103854262A (en) * | 2014-03-20 | 2014-06-11 | 西安电子科技大学 | Medical image noise reduction method based on structure clustering and sparse dictionary learning |
| CN104063886A (en) * | 2014-03-24 | 2014-09-24 | 杭州电子科技大学 | Nuclear magnetic resonance image reconstruction method based on sparse representation and non-local similarity |
| CN104657951A (en) * | 2015-03-02 | 2015-05-27 | 桂林电子科技大学 | Multiplicative noise removal method for image |
| CN105513026A (en) * | 2015-12-17 | 2016-04-20 | 浙江大学 | Compressed sensing reconstruction method based on image nonlocal similarity |
Non-Patent Citations (1)
| Title |
|---|
| 赵井坤 等: "基于稀疏表示与非局部相似的图像去噪算法", 《计算机应用》 * |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107085839A (en) * | 2017-06-14 | 2017-08-22 | 西安电子科技大学 | SAR image method for reducing speckle with sparse coding is strengthened based on texture |
| CN107085839B (en) * | 2017-06-14 | 2020-01-07 | 西安电子科技大学 | SAR image speckle reduction method based on texture enhancement and sparse coding |
| CN107358589B (en) * | 2017-07-17 | 2019-11-26 | 桂林电子科技大学 | A kind of denoising method of combination histogram of gradients and low-rank constraint |
| CN107358589A (en) * | 2017-07-17 | 2017-11-17 | 桂林电子科技大学 | A kind of combination histogram of gradients and the denoising method of low-rank constraint |
| CN107730468B (en) * | 2017-10-23 | 2020-11-10 | 鹰艾思科技(深圳)有限公司 | Method for recovering sharp image under unmanned aerial vehicle fuzzy noise image |
| CN107730468A (en) * | 2017-10-23 | 2018-02-23 | 鹰艾思科技(深圳)有限公司 | The restoration methods of picture rich in detail under a kind of UAV Fuzzy noise image |
| CN108765322A (en) * | 2018-05-16 | 2018-11-06 | 上饶师范学院 | Image de-noising method and device |
| CN108765322B (en) * | 2018-05-16 | 2021-04-27 | 上饶师范学院 | Image denoising method and device |
| CN108710150A (en) * | 2018-05-22 | 2018-10-26 | 中国海洋石油集团有限公司 | A kind of earthquake random noise minimizing technology based on steady singular spectrum analysis |
| CN111402173A (en) * | 2020-03-24 | 2020-07-10 | 北京工业大学 | Hybrid noise removing method and device, electronic equipment and storage medium |
| CN111402173B (en) * | 2020-03-24 | 2024-04-16 | 北京工业大学 | Mixed noise removing method and device, electronic equipment and storage medium |
| CN111586262A (en) * | 2020-03-27 | 2020-08-25 | 成都航天通信设备有限责任公司 | Image denoising method based on field army image transmission |
| CN113192078A (en) * | 2021-04-19 | 2021-07-30 | 清华大学 | Rapid sea surface SAR image superpixel segmentation method based on density |
| CN113192078B (en) * | 2021-04-19 | 2022-12-02 | 清华大学 | Rapid sea surface SAR image superpixel segmentation method based on density |
| CN113487522A (en) * | 2021-09-08 | 2021-10-08 | 深圳市诚识科技有限公司 | Multi-channel switching noise reduction method for image communication |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106204482B (en) | 2019-03-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106204482B (en) | Based on the mixed noise minimizing technology that weighting is sparse | |
| CN114140353B (en) | Swin-Transformer image denoising method and system based on channel attention | |
| CN110599409B (en) | Convolutional neural network image denoising method based on multi-scale convolutional groups and parallel | |
| CN109785258B (en) | Face image restoration method based on multi-discriminator generated countermeasure network | |
| Zhang et al. | Learning deep CNN denoiser prior for image restoration | |
| CN110796625B (en) | Image compressed sensing reconstruction method based on group sparse representation and weighted total variation | |
| WO2022267641A1 (en) | Image defogging method and system based on cyclic generative adversarial network | |
| CN109035163A (en) | A kind of adaptive denoising method based on deep learning | |
| CN107464226B (en) | A kind of image de-noising method based on improvement two-dimensional empirical mode decomposition algorithm | |
| CN110232661A (en) | Low illumination colour-image reinforcing method based on Retinex and convolutional neural networks | |
| CN110163815A (en) | Low-light (level) restoring method based on multistage variation self-encoding encoder | |
| CN107680077A (en) | A kind of non-reference picture quality appraisement method based on multistage Gradient Features | |
| CN109493295B (en) | Non-local haar transform image denoising method | |
| CN112270654A (en) | Image denoising method based on multi-channel GAN | |
| Chen et al. | Densely connected convolutional neural network for multi-purpose image forensics under anti-forensic attacks | |
| Choi et al. | Optimal combination of image denoisers | |
| CN110490816A (en) | A kind of underwater Heterogeneous Information data noise reduction | |
| CN106972862A (en) | Based on the sparse compressed sensing image reconstructing method of group for blocking nuclear norm minimum | |
| CN104732498B (en) | A kind of thresholded image denoising method based on non-downsampling Contourlet conversion | |
| CN111832517A (en) | Low-definition face key point detection method based on gated convolution | |
| CN106296583B (en) | Based on image block group sparse coding and the noisy high spectrum image ultra-resolution ratio reconstructing method that in pairs maps | |
| CN106169180A (en) | A kind of non-local sparse based on group represents additive noise minimizing technology | |
| CN107292855A (en) | A kind of image de-noising method of the non local sample of combining adaptive and low-rank | |
| Hu et al. | Patch-based low-rank minimization for image denoising | |
| CN104050644A (en) | SAR image denoising method based on non-local restriction sparse representation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |