CN109272539A - The decomposition method of image texture and structure based on guidance figure Total Variation - Google Patents
The decomposition method of image texture and structure based on guidance figure Total Variation Download PDFInfo
- Publication number
- CN109272539A CN109272539A CN201811066262.2A CN201811066262A CN109272539A CN 109272539 A CN109272539 A CN 109272539A CN 201811066262 A CN201811066262 A CN 201811066262A CN 109272539 A CN109272539 A CN 109272539A
- Authority
- CN
- China
- Prior art keywords
- image
- texture
- pixel
- guidance
- partial structurtes
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 56
- 238000000354 decomposition reaction Methods 0.000 title claims abstract description 46
- 238000001914 filtration Methods 0.000 claims abstract description 21
- 230000002146 bilateral effect Effects 0.000 claims description 9
- 238000012937 correction Methods 0.000 claims description 9
- 230000008569 process Effects 0.000 claims description 9
- 238000004364 calculation method Methods 0.000 claims description 5
- 239000013598 vector Substances 0.000 claims description 3
- 230000000694 effects Effects 0.000 abstract description 9
- 238000005457 optimization Methods 0.000 abstract description 4
- 238000009499 grossing Methods 0.000 abstract description 3
- 238000012545 processing Methods 0.000 abstract description 2
- 238000000605 extraction Methods 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 238000013459 approach Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 241000208340 Araliaceae Species 0.000 description 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 1
- 235000003140 Panax quinquefolius Nutrition 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000007717 exclusion Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 235000008434 ginseng Nutrition 0.000 description 1
- 230000002401 inhibitory effect Effects 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 230000031068 symbiosis, encompassing mutualism through parasitism Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/40—Analysis of texture
-
- 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
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20024—Filtering details
- G06T2207/20028—Bilateral filtering
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Processing (AREA)
Abstract
The present invention discloses the decomposition method of a kind of image texture based on guidance figure Total Variation and structure, belongs to technical field of image processing.The decomposition method of image texture and structure provided by the invention based on guidance figure Total Variation, the local primary structure of filtering reconstructed image is first schemed based on guidance, then Texture descriptor is calculated according to the local structural graph of reconstruct, multiple dimensioned Total Variation and block shift method is finally combined to improve the accuracy and computational efficiency of texture and STRUCTURE DECOMPOSITION.Technical solution of the present invention can obtain better texture/STRUCTURE DECOMPOSITION effect to the biggish image of noise, structure and texture to different scale in image, which also can be carried out, accurately decomposes, the structure sheaf decomposited is able to maintain the original light and shade variation of image, avoids structural fuzzy caused by single use local smoothing method or global optimization method bring color lump effect.In addition, the feature extracted required for technical solution of the present invention is simple, independent of the study to great amount of images sample.
Description
Technical field
The invention belongs to technical field of image processing, specifically, being related to a kind of figure based on guidance figure Total Variation
As the decomposition method of texture and structure.
Background technique
Texture and structure are the intrinsic most basic visual signatures of image, and the texture and structure in separate picture are to further
Handle and understand image important in inhibiting.Texture is accurately and efficiently separated in complex scene image and structure is one has
The work of challenge.In practical applications, the method based on filtering and be the common texture knot of two classes based on full variational approach
Structure decomposition method.
It completes to decompose by calculating the difference between image local texture and structure feature based on the method for filtering, advantage
It is that method is simple, intuitive, speed is fast, and defect is to be difficult to accurate reconstruction to weak edge, is easy to cause parts of images important feature quilt
It is excessively fuzzy, it is weaker to the stronger texture capacity of decomposition of localized variation.Total Variation can preferably express texture/STRUCTURE DECOMPOSITION
Problem, the data item or regular terms that the texture and structure of image can be different from model are corresponding, by minimizing energy function
Globally optimal solution can be obtained, model also has higher controllability.But texture and STRUCTURE DECOMPOSITION based on full variation are to texture
Description method is very sensitive, and the selection for surpassing ginseng is also very difficult, and global optimization is easy to cause the structural images after decomposing excessive
Smoothly, lack reasonable local light and shade variation, can degenerate in partial region for single color lump.In addition, based on Total Variation point
It solves texture and Structure Calculation amount is big, have the application of strict demand to the calculating time and be not suitable for.
Summary of the invention
Technical problem to be solved by the invention is to provide a kind of image textures and knot based on guidance figure Total Variation
The decomposition method of structure, to improve the accuracy and computational efficiency of image texture and STRUCTURE DECOMPOSITION.
The invention discloses the decomposition methods of a kind of image texture based on guidance figure Total Variation and structure, including with
Lower step:
S1, original image I to be decomposed and guidance figure are inputted into partial structurtes reconstructed module；
S2, schemed based on original image I to be decomposed and guidance, gaussian filtering, joint bilateral filtering used by successive ignition,
Reconstruct partial structurtes image；
S3, the textural characteristics F for extracting partial structurtes imagep, FpFor 5 dimensional feature vectors,WhereinRespectively pixel p is three in Lab color space
Value on channel,Respectively indicate single order and second dervative of the pixel p on the channel L；
S4, the textural characteristics F for calculating partial structurtes imagepRegion covariance, obtain partial structurtes image texture
Tp, wherein participating in the tile size that region covariance calculates is k, and the initial value of k is 15；
S5, strategy is translated to the texture T of the partial structurtes image of acquisition using blockpIt is corrected；
S6, the texture of the partial structurtes image after original image and correction is inputted into full variation texture/STRUCTURE DECOMPOSITION module,
Decomposite the global structure layer of image；
S7, partial structurtes reconstructed module is input to using the structure sheaf S* after decomposition as new guidance figure；
S8, the tile size k in S4 step is updated are as follows:
S9, S2~S8 step 4 time is repeated, the S* that last time iteration is obtained is exported as the structure sheaf S of original image；
S10, finally obtained original image structure sheaf S in S9 step is subtracted with original image I, obtains the texture layer T of original image.
Method as described above, wherein the realization process of S2 step specifically includes:
S21, initial space scale parameter σ is set according to texture sizes, σsInitial value be 3；S22, according to σsTo drawing
It leads figure and carries out gaussian filtering, smoothed image grain details；
S23, joint bilateral filtering is carried out to original image I and guidance figure, wherein color weight drawing after the calculating of S21 step
It leads and is obtained on figure；
S24, scale parameter σ is updateds=2 σs+ 1, and using the filtered image of S23 step as new guidance figure；
S25, step 3~4 time S22~S24 are repeated, last time executes the bilateral filtering result that S23 step obtains and is
Partial structurtes image after reconstruct.
Method as described above, wherein the realization process of S5 step specifically includes:
S51, in the partial structurtes image of acquisition, find out all image blocks comprising pixel p, and be labeled as candidate image
Block Ωq, wherein q is the center pixel of candidate image block, and meets q ∈ Ωp；
S52, the texture difference value D (Ω for calculating each candidate image blockq), calculation formula is as follows:
In formula, Dx(Tr) and Dy(Tr) respectively indicate candidate image block ΩqThe texture of middle pixel r is on horizontal and vertical
Difference；
S53, the smallest candidate image block of texture difference value is chosen, and with the texture T of its center pixelqReplacement pixel
The texture of the image block of p.
Method as described above, wherein the realization process of S6 step specifically includes:
S61, the weight w that regularization term in Total Variation is calculated according to the texture of the partial structurtes image after correctionx,p
And wy,p, wx,pIndicate weight of the pixel p in the horizontal direction on x, wy,pIt indicates weight of the pixel p on vertical direction y, calculates public
Formula is as follows:
In formula, Dx(Tp) indicate the difference of the texture after pixel p corrects in the x direction,It indicates to original image
Luminance component seek single order local derviation in the x direction；Dy(Tp) indicate the difference of the texture after pixel p corrects in y-direction,Expression seeks single order local derviation to the luminance component of original image in y-direction；εt=εs=1e-6；
S62, pass through the following full Variational Decomposition model of solution, the structure sheaf S* after obtaining decomposition:
In formula, SpFor value of the pixel p in required structure tomographic image, IpIt is pixel p in the original image that user inputs
Value,The single order local derviation for being pixel p in structure tomographic image on the direction x,For pixel p in structure tomographic image the side y
Upward single order local derviation, weighting factor of the λ between fidelity term and regularization term, λ ∈ [0.01,0.05].
The decomposition method of image texture and structure provided by the invention based on guidance figure Total Variation has following excellent
Point:
1, better texture/STRUCTURE DECOMPOSITION effect can be obtained to the biggish image of noise.
2, the structure to different scale in image and texture, which also can be carried out, accurately decomposes.
3, the structure sheaf decomposited is able to maintain the original light and shade variation of image, avoids caused by single use local smoothing method
Structural fuzzy or global optimization method bring color lump effect.
4, the feature of required extraction is simple, independent of the study to great amount of images sample.
Detailed description of the invention
The drawings described herein are used to provide a further understanding of the present invention, constitutes a part of the invention, this hair
Bright illustrative embodiments and their description are used to explain the present invention, and are not constituted improper limitations of the present invention.In the accompanying drawings:
Fig. 1 is the stream of the decomposition method embodiment one the present invention is based on the image texture of guidance figure Total Variation and structure
Cheng Tu；
Fig. 2 is the stream of the decomposition method embodiment two the present invention is based on the image texture of guidance figure Total Variation and structure
Cheng Tu.
Specific embodiment
Carry out the embodiment that the present invention will be described in detail below in conjunction with examples and drawings, how the present invention is applied whereby
Technological means solves technical problem and reaches the realization process of technical effect to fully understand and implement.
Lack light and shade variation and decomposition method based on filtering in smooth grain region for traditional full Variational Solution Used method
It is be easy to cause excessively fuzzy defect, the present invention provides point of a kind of image texture based on guidance figure Total Variation and structure
Solution method optimizes two layers of iteration structure by Partial Reconstruction and full variation and improves image texture/STRUCTURE DECOMPOSITION accuracy.Fig. 1
For the present invention is based on the flow chart of the image texture of guidance figure Total Variation and the decomposition method embodiment one of structure, Fig. 2 is
The present invention is based on the flow charts of the image texture of guidance figure Total Variation and the decomposition method embodiment two of structure.With reference to Fig. 1
With shown in Fig. 2, the decomposition method of the image texture and structure of the invention based on guidance figure Total Variation may include following
Step (S1~S10):
S1, original image I to be decomposed and guidance figure are inputted into partial structurtes reconstructed module.
The copy that guidance figure described in the step is original image I to be decomposed.
S2, schemed based on original image I to be decomposed and guidance, gaussian filtering, joint bilateral filtering used by successive ignition,
Reconstruct partial structurtes image.
Specifically, partial structurtes reconstructed module, which is based on guidance figure, carries out successive ignition filtering (internal layer iteration), image is removed
Tiny texture and noise, reconstruct partial structurtes image.Restructuring procedure is specially (S21~S25):
S21, initial space scale parameter σ is set according to texture sizes, σsInitial value be 3.It is set according to texture size
Set initial space scale parameter σs, σ under default situationssValue be initial value, i.e. σs=3.It in a particular application, can be to σs
Initial setting be modified.
S22, according to σsGuidance is schemed to carry out gaussian filtering, smoothed image grain details.
S23, joint bilateral filtering is carried out to original image I and guidance figure, wherein color weight drawing after the calculating of S21 step
It leads and is obtained on figure.
S24, scale parameter σ is updateds=2 σs+ 1, and using the filtered image of S23 step as new guidance figure.
S25, step 3~4 time S22~S24 are repeated, last time executes the bilateral filtering result that S23 step obtains and is
Partial structurtes image after reconstruct.
S3, the textural characteristics F for extracting partial structurtes imagep, FpFor 5 dimensional feature vectors,WhereinRespectively pixel p is three in Lab color space
Value on channel,Respectively indicate single order and second dervative of the pixel p on the channel L.
In the concrete application of this feature extraction step, the extracting method of textural characteristics can be replaced, both be can be used
Deterministic Texture Segmentation Algorithm, such as partial binary operator (LBP) or gray scale symbiosis, can also be based on the side of study
Method obtains textural characteristics and indicates, the textural characteristics that the method such as based on deep learning obtains specific type image indicate, but are based on
The method of study usually requires the largely sample data with structure or texture markings.Therefore the method for the invention in practical application
When, feature extracting method can be replaced according to actual needs, remaining step is without change.
S4, the textural characteristics F for calculating partial structurtes imagepRegion covariance, obtain partial structurtes image texture
Tp, wherein participating in the tile size that region covariance calculates is k, and the initial value of k is 15.
The region covariance of textural characteristics is calculated to characterize the degree that pixel belongs to texture region, obtains texture Tp,
The middle initial pictures block size k for participating in region covariance and calculating, the value of k is initial setting 15 under default situations, is specifically being answered
In, the initial setting of k can be modified.
S5, strategy is translated to the texture T of the partial structurtes image of acquisition using blockpIt is corrected.
Texture figure is inputted into texture correction module, the accuracy of texture estimation is further increased by the module.
Correction course is specially (S51~S53):
S51, in the partial structurtes image of acquisition, find out all image blocks comprising pixel p, and be labeled as candidate image
Block Ωq, wherein q is the center pixel of candidate image block, and meets q ∈ Ωp。
S52, the texture difference value D (Ω for calculating each candidate image blockq), calculation formula is as follows:
In formula, Dx(Tr) and Dy(Tr) respectively indicate candidate image block ΩqThe texture of middle pixel r is on horizontal and vertical
Difference.
S53, the smallest candidate image block of texture difference value is chosen, and with the texture T of its center pixelqReplacement pixel
The texture of the image block of p.
In the step, by choosing the smallest candidate image block of texture difference value, and with the texture of its center pixel
TqThe texture of the image block of replacement pixel p so far completes the texture correction translated based on block.
S6, the texture of the partial structurtes image after original image and correction is inputted into full variation texture/STRUCTURE DECOMPOSITION module,
Decomposite the global structure layer of image.
The decomposable process of global structure layer is specially (S61~S62):
S61, the weight w that regularization term in Total Variation is calculated according to the texture of the partial structurtes image after correctionx,p
And wy,p, wx,pIndicate weight of the pixel p in the horizontal direction on x, wy,pIt indicates weight of the pixel p on vertical direction y, calculates public
Formula is as follows:
In formula, Dx(Tp) indicate the difference of the texture after pixel p corrects in the x direction,It indicates to original image
Luminance component seek single order local derviation in the x direction；Dy(Tp) indicate the difference of the texture after pixel p corrects in y-direction,Expression seeks single order local derviation to the luminance component of original image in y-direction；εt=εs=1e-6.
εtAnd εsFor minimum positive number, for avoid denominator from being 0 and caused by system it is unstable, usually take εt=εs=1e-
6。
S62, pass through the following full Variational Decomposition model of solution, the structure sheaf S* after obtaining decomposition:
In formula, SpFor value of the pixel p in required structure tomographic image, IpIt is pixel p in the original image that user inputs
Value,The single order local derviation for being pixel p in structure tomographic image on the direction x,For pixel p in structure tomographic image the side y
Upward single order local derviation, weighting factor of the λ between fidelity term and regularization term, λ ∈ [0.01,0.05].The preferred value of λ is
0.02, usually 0.02, i.e. λ=0.02 under default situations are set by the initial value of λ.
S7, partial structurtes reconstructed module is input to using the structure sheaf S* after decomposition as new guidance figure.
S8, the tile size k in S4 step is updated are as follows:
In the step, by the tile size reduced in S4 step, the scale of Image Iterative is reduced.
S9, S2~S8 step 4 time is repeated, the S* that last time iteration is obtained is exported as the structure sheaf S of original image.
S10, finally obtained original image structure sheaf S in S9 step is subtracted with original image I, obtains the texture layer T of original image.
The decomposition method of image texture and structure compared to the prior art, technical solution provided by the invention have following
Advantage:
One, better texture/STRUCTURE DECOMPOSITION effect can be obtained to the biggish image of noise.
Before actually decomposing texture and structure, figure is not only effectively reduced by the reconstruct for carrying out partial structurtes to image
Noise as in, and the major side of image can be kept well, it is provided for the estimation of subsequent texture and decomposing module
The input of high quality, therefore accurate texture/STRUCTURE DECOMPOSITION effect can be obtained to strong noise image.
Two, the structure to different scale in image and texture, which also can be carried out, accurately decomposes.
By introducing multiple dimensioned iteration, skill provided by the invention in partial structurtes reconstructed module and full Variational Decomposition module
Art scheme can not only decompose texture and structure on different scale, and two layers iteration structure (partial structurtes reconstruct and
Full Variational Decomposition module respectively corresponds internal layer and external iteration) the similar texture of height and knot on scale can be better discriminated between
Structure, and in decomposable process, it is translated by introduce region covariance and block, texture and structure can not only be more accurately distinguished between,
And energy effective position major side position, to further enhance the accuracy of texture and STRUCTURE DECOMPOSITION.
Three, the structure sheaf decomposited is able to maintain the original light and shade variation of image, and avoiding single use local smoothing method causes
Structural fuzzy or global optimization method bring color lump effect.
Partial Reconstruction module in technical solution provided by the invention can preferably keep structural images in the light and shade of part
Variation, and the decomposition based on full variation then can effectively inhibit excess smoothness caused by partial approach, to be effectively combined office
The advantage in portion and global decomposition method.
Four, the feature of required extraction is simple, independent of the study to great amount of images sample.
Characteristic extraction step in technical solution provided by the invention is simple and easy, is not necessarily based on great amount of images sample learning
Character representation, this realize the estimation of the texture in the present invention can not only efficiently, and can be according to practical application request spirit
Extension living.
In conclusion the decomposition method of the image texture and structure provided by the invention based on guidance figure Total Variation,
The local primary structure that filtering reconstructed image is first schemed based on guidance, then calculates texture description according to the local structural graph of reconstruct
Symbol finally combines multiple dimensioned Total Variation and block shift method to improve the accuracy of texture and STRUCTURE DECOMPOSITION and calculates effect
Rate.
Several preferred embodiments of the invention have shown and described in above description, but as previously described, it should be understood that the present invention
Be not limited to forms disclosed herein, should not be regarded as an exclusion of other examples, and can be used for various other combinations,
Modification, and can be in contemplated scope of the present invention, modifications can be made through the above teachings or related fields of technology or knowledge.And this
The modifications and changes that field personnel are carried out do not depart from the spirit and scope of the present invention, then all should be in appended claims of the present invention
Protection scope in.
Claims (4)
1. a kind of decomposition method of image texture and structure based on guidance figure Total Variation, which is characterized in that including following
Step:
S1, original image I to be decomposed and guidance figure are inputted into partial structurtes reconstructed module；
S2, it is reconstructed based on original image I to be decomposed and guidance figure by successive ignition using gaussian filtering, joint bilateral filtering
Local structural images out；
S3, the textural characteristics F for extracting partial structurtes imagep, FpFor 5 dimensional feature vectors,WhereinRespectively pixel p is three in Lab color space
Value on channel,Respectively indicate single order and second dervative of the pixel p on the channel L；
S4, the textural characteristics F for calculating partial structurtes imagepRegion covariance, obtain partial structurtes image texture Tp,
In, participating in the tile size that region covariance calculates is k, and the initial value of k is 15；
S5, strategy is translated to the texture T of the partial structurtes image of acquisition using blockpIt is corrected；
S6, the texture of the partial structurtes image after original image and correction is inputted into full variation texture/STRUCTURE DECOMPOSITION module, decomposed
The global structure layer of image out；
S7, partial structurtes reconstructed module is input to using the structure sheaf S* after decomposition as new guidance figure；
S8, the tile size k in S4 step is updated are as follows:S9, S2~S8 step 4 time is repeated, by last
The S* that secondary iteration obtains is exported as the structure sheaf S of original image；
S10, finally obtained original image structure sheaf S in S9 step is subtracted with original image I, obtains the texture layer T of original image.
2. the method as described in claim 1, which is characterized in that the realization process of S2 step specifically includes:
S21, initial space scale parameter σ is set according to texture sizes, σsInitial value be 3；
S22, according to σsGuidance is schemed to carry out gaussian filtering, smoothed image grain details；
S23, joint bilateral filtering is carried out to original image I and guidance figure, wherein guidance figure of the color weight after the calculating of S21 step
Upper acquisition；
S24, scale parameter σ is updateds=2 σs+ 1, and using the filtered image of S23 step as new guidance figure；
S25, step 3~4 time S22~S24 are repeated, it is to reconstruct that last time, which executes the bilateral filtering result that S23 step obtains,
Partial structurtes image afterwards.
3. method according to claim 2, which is characterized in that the realization process of S5 step specifically includes:
S51, in the partial structurtes image of acquisition, find out all image blocks comprising pixel p, and be labeled as candidate image block
Ωq, wherein q is the center pixel of candidate image block, and meets q ∈ Ωp；
S52, the texture difference value D (Ω for calculating each candidate image blockq), calculation formula is as follows:
In formula, Dx(Tr) and Dy(Tr) respectively indicate candidate image block ΩqDifference of the texture of middle pixel r on horizontal and vertical
Point；
S53, the smallest candidate image block of texture difference value is chosen, and with the texture T of its center pixelqThe figure of replacement pixel p
As the texture of block.
4. method as claimed in claim 3, which is characterized in that the realization process of S6 step specifically includes:
S61, the weight w that regularization term in Total Variation is calculated according to the texture of the partial structurtes image after correctionx,pWith
wy,p, wx,pIndicate weight of the pixel p in the horizontal direction on x, wy,pIndicate weight of the pixel p on vertical direction y, calculation formula
It is as follows:
In formula, Dx(Tp) indicate the difference of the texture after pixel p corrects in the x direction,Indicate the brightness to original image
Component seeks single order local derviation in the x direction；Dy(Tp) indicate the difference of the texture after pixel p corrects in y-direction,Table
Show and single order local derviation is asked in y-direction to the luminance component of original image；εt=εs=1e-6；
S62, pass through the following full Variational Decomposition model of solution, the structure sheaf S* after obtaining decomposition:
In formula, SpFor value of the pixel p in required structure tomographic image, IpFor value of the pixel p in the original image that user inputs,The single order local derviation for being pixel p in structure tomographic image on the direction x,For pixel p in structure tomographic image the direction y
On single order local derviation, weighting factor of the λ between fidelity term and regularization term, λ ∈ [0.01,0.05].
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811066262.2A CN109272539A (en) | 2018-09-13 | 2018-09-13 | The decomposition method of image texture and structure based on guidance figure Total Variation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811066262.2A CN109272539A (en) | 2018-09-13 | 2018-09-13 | The decomposition method of image texture and structure based on guidance figure Total Variation |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109272539A true CN109272539A (en) | 2019-01-25 |
Family
ID=65188899
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811066262.2A Pending CN109272539A (en) | 2018-09-13 | 2018-09-13 | The decomposition method of image texture and structure based on guidance figure Total Variation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109272539A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110599408A (en) * | 2019-07-25 | 2019-12-20 | 安庆师范大学 | Region selective multi-scale de-texturing method based on image texture |
| CN110895693A (en) * | 2019-09-12 | 2020-03-20 | 华中科技大学 | Authentication method and authentication system for anti-counterfeiting information of certificate |
| CN111369510A (en) * | 2020-02-28 | 2020-07-03 | 四川大学华西医院 | Method for automatically estimating choroid thickness |
| CN112862715A (en) * | 2021-02-08 | 2021-05-28 | 天津大学 | Real-time and controllable scale space filtering method |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103700117A (en) * | 2013-11-21 | 2014-04-02 | 北京工业大学 | Robust optical flow field estimating method based on TV-L1 variation model |
| CN105469407A (en) * | 2015-11-30 | 2016-04-06 | 华南理工大学 | Facial image layer decomposition method based on improved guide filter |
| CN105607122A (en) * | 2015-12-23 | 2016-05-25 | 西南科技大学 | Seismic texture extraction and enhancement method based on total variation seismic data decomposition model |
| KR101635309B1 (en) * | 2015-01-14 | 2016-07-01 | 포항공과대학교 산학협력단 | Apparatus and method of textrue filtering using patch shift |
| CN106683129A (en) * | 2016-12-13 | 2017-05-17 | 中山大学 | Method for decomposing relatively reduced texture based on bilateral kernel regression and device thereof |
| CN107798665A (en) * | 2017-11-07 | 2018-03-13 | 天津大学 | Underwater picture Enhancement Method based on structural texture layering |
-
2018
- 2018-09-13 CN CN201811066262.2A patent/CN109272539A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103700117A (en) * | 2013-11-21 | 2014-04-02 | 北京工业大学 | Robust optical flow field estimating method based on TV-L1 variation model |
| KR101635309B1 (en) * | 2015-01-14 | 2016-07-01 | 포항공과대학교 산학협력단 | Apparatus and method of textrue filtering using patch shift |
| CN105469407A (en) * | 2015-11-30 | 2016-04-06 | 华南理工大学 | Facial image layer decomposition method based on improved guide filter |
| CN105607122A (en) * | 2015-12-23 | 2016-05-25 | 西南科技大学 | Seismic texture extraction and enhancement method based on total variation seismic data decomposition model |
| CN106683129A (en) * | 2016-12-13 | 2017-05-17 | 中山大学 | Method for decomposing relatively reduced texture based on bilateral kernel regression and device thereof |
| CN107798665A (en) * | 2017-11-07 | 2018-03-13 | 天津大学 | Underwater picture Enhancement Method based on structural texture layering |
Non-Patent Citations (3)
| Title |
|---|
| HAO WU 等: "Region covariance based total variation optimization for structure-texture decomposition", 《MULTIMEDIA TOOLS AND APPLICATIONS》 * |
| QI ZHANG 等: "Rolling Guidance Filter", 《COMPUTER VISION – ECCV 2014》 * |
| 芦碧波 等: "各向异性全变分引导滤波及其Split Bregman方法", 《激光与光电子学进展》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110599408A (en) * | 2019-07-25 | 2019-12-20 | 安庆师范大学 | Region selective multi-scale de-texturing method based on image texture |
| CN110895693A (en) * | 2019-09-12 | 2020-03-20 | 华中科技大学 | Authentication method and authentication system for anti-counterfeiting information of certificate |
| CN110895693B (en) * | 2019-09-12 | 2022-04-26 | 华中科技大学 | Authentication method and authentication system for anti-counterfeiting information of certificate |
| CN111369510A (en) * | 2020-02-28 | 2020-07-03 | 四川大学华西医院 | Method for automatically estimating choroid thickness |
| CN111369510B (en) * | 2020-02-28 | 2022-07-01 | 四川大学华西医院 | Method for automatically estimating choroid thickness |
| CN112862715A (en) * | 2021-02-08 | 2021-05-28 | 天津大学 | Real-time and controllable scale space filtering method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Li et al. | Infrared and visible image fusion using a deep learning framework | |
| CN110992262B (en) | Remote sensing image super-resolution reconstruction method based on generation countermeasure network | |
| CN109272539A (en) | The decomposition method of image texture and structure based on guidance figure Total Variation | |
| CN108875935B (en) | Natural image target material visual characteristic mapping method based on generation countermeasure network | |
| CN104899845B (en) | A kind of more exposure image fusion methods based on the migration of l α β spatial scenes | |
| CN105631807B (en) | The single-frame image super-resolution reconstruction method chosen based on sparse domain | |
| CN107066916B (en) | Scene semantic segmentation method based on deconvolution neural network | |
| CN106204447A (en) | The super resolution ratio reconstruction method with convolutional neural networks is divided based on total variance | |
| CN109711401B (en) | Text detection method in natural scene image based on Faster Rcnn | |
| CN107133496B (en) | Gene feature extraction method based on manifold learning and closed-loop deep convolution double-network model | |
| CN110570377A (en) | group normalization-based rapid image style migration method | |
| CN110634147A (en) | Image matting method based on bilateral boot up-sampling | |
| CN107507146B (en) | Natural image soft shadow elimination method | |
| CN102915527A (en) | Face image super-resolution reconstruction method based on morphological component analysis | |
| CN110276753A (en) | Objective self-adapting hidden method based on the mapping of feature space statistical information | |
| CN109064525A (en) | Picture format conversion method, device, equipment and storage medium | |
| CN109191377A (en) | A kind of image magnification method based on interpolation | |
| CN113052859A (en) | Super-pixel segmentation method based on self-adaptive seed point density clustering | |
| CN115641583B (en) | Point cloud detection method, system and medium based on self-supervision and active learning | |
| CN112037109A (en) | Improved image watermarking method and system based on saliency target detection | |
| CN112686830A (en) | Super-resolution method of single depth map based on image decomposition | |
| CN109741258B (en) | Image super-resolution method based on reconstruction | |
| CN112766340A (en) | Depth capsule network image classification method and system based on adaptive spatial mode | |
| CN114897694A (en) | Image super-resolution reconstruction method based on mixed attention and double-layer supervision | |
| CN109063725A (en) | More figure regularization matrix of depths decomposition methods towards multiple view cluster |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190125 |