CN106875359B - A kind of sample block image repair method based on layering boot policy - Google Patents
A kind of sample block image repair method based on layering boot policy Download PDFInfo
- Publication number
- CN106875359B CN106875359B CN201710082887.7A CN201710082887A CN106875359B CN 106875359 B CN106875359 B CN 106875359B CN 201710082887 A CN201710082887 A CN 201710082887A CN 106875359 B CN106875359 B CN 106875359B
- Authority
- CN
- China
- Prior art keywords
- image
- top layer
- sample block
- information
- repaired
- 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.)
- Active
Links
- 230000008439 repair process Effects 0.000 title claims abstract description 44
- 238000000034 method Methods 0.000 title claims abstract description 36
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 22
- 238000005070 sampling Methods 0.000 claims abstract description 10
- 238000004040 coloring Methods 0.000 claims description 4
- 238000005267 amalgamation Methods 0.000 abstract description 4
- 230000000007 visual effect Effects 0.000 abstract description 4
- 230000000694 effects Effects 0.000 abstract description 3
- 238000010586 diagram Methods 0.000 description 3
- 150000001875 compounds Chemical class 0.000 description 2
- 230000015556 catabolic process Effects 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification 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
- G06T5/77—Retouching; Inpainting; Scratch removal
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/50—Image enhancement or restoration using two or more images, e.g. averaging or subtraction
-
- 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
- G06T2207/20021—Dividing image into blocks, subimages or windows
-
- 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
- G06T2207/20212—Image combination
- G06T2207/20221—Image fusion; Image merging
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Processing Or Creating Images (AREA)
- Editing Of Facsimile Originals (AREA)
Abstract
The present invention discloses a kind of sample block image repair method based on layering boot policy, including carrying out gaussian pyramid decomposition to original image to be repaired, obtain the decomposition image of multilayer different resolution, the decomposition image is successively repaired according to the sequence from top layer to bottom, first sample block restorative procedure is based on to the decomposition image of top layer to repair, it obtains the top layer images repaired and repairs result, result up-sampling treatment is repaired to the top layer images, image is decomposed with secondary top layer again to merge, and the information in region to be filled in amalgamation result is repaired into time top layer as guidance information and decomposes image, repeat this process, it repairs and finishes until all layers of regions to be filled, a kind of sample block image repair method based on layering boot policy of the invention, it is with applied widely, repairing effect is good and practical feature , it can be applicable in the image repair of large area breakage, meet visual effect.
Description
Technical field
The present invention relates to image restoration technology fields, more particularly to a kind of sample block image based on layering boot policy
Restorative procedure.
Background technique
Image repair is an actual application problem, be mainly used in object hide, historical relic loss repair, photo it is above
Word or the removal on date etc..
Current most representative restorative procedure has two classes: one kind is the image repair method based on variation partial differential；Separately
One kind is the technology based on sample block.When existing sample block restorative procedure repairs the image damaged area of large area, exist
The problem of being difficult to keep structural integrity.
Summary of the invention
The present invention is completed to solve in the prior art insufficient, and the object of the present invention is to provide a kind of applicable models
Enclose wide, repairing effect is good with practical feature, can be applicable in the image repair of large area breakage, meet visual effect based on
It is layered the sample block image repair method of boot policy.
A kind of sample block image repair method based on layering boot policy of the invention, including Step 1: to be repaired
Original image carry out gaussian pyramid decomposition, obtain multilayer different resolution decomposition image；Step 2: according to from top layer to
The sequence of bottom successively repairs the decomposition image, is first carried out to the decomposition image of top layer based on sample block restorative procedure
It repairs, obtains the top layer images repaired and repair result；Step 3: result up-sampling treatment is repaired to the top layer images, then
Image is decomposed with secondary top layer to merge, and repairs time top for the information in region to be filled in amalgamation result as guidance information
Layer decomposes image, repeats this process, repairs and finishes until all layers of regions to be filled.
A kind of sample block image repair method based on layering boot policy of the invention may also is that
If filling forward positionOn have a point p, in the to be filled piece of Ψ centered on ppIn there are two region: Given information
Region ΨpaWith region Ψ to be filledpb, the restorative procedure based on sample block utilizes ΨpMiddle ΨpaInformation, according to formula (1) count
Calculation searches for Best exemplar patch in sample resource Φ
Wherein, Φ indicates the sample areas in image,It is and ΨpApart from the smallest sample block, d (Ψpa, Ψpa) table
Show ΨpaAnd ΨqaBetween color component and gradient component Euclidean distance；
Ψ is calculated further according to formula (2)pAnd ΨqBetween color component and gradient component Euclidean distance,
Wherein, Ip, IqIt is corresponding RGB vector,Indicate the gradient vector of image；
If top layer images GKS is obtained by up-sampling treatment after reparationK-1, SK-1With GK-1Merging obtains AK-1, at this moment AK-1In
Ω in information filled, using boot policy sample block repair image when, ΨpbIn information as guidance
Information, if to be filled piece is Ψp, Best exemplar patch is found at this timeBest exemplar patch is calculated using formula (3)
Wherein, ΨpaIt is Given information region, ΨpbIt is region to be filled, information therein is guidance information；
Distance is calculated using the colouring information of sample block by formula (4) again,
Wherein, Ip, IqIt is corresponding RGB vector.
A kind of sample block image repair method based on layering boot policy of the invention, including to original graph to be repaired
As carrying out gaussian pyramid decomposition, the decomposition image of multilayer different resolution is obtained；According to the sequence from top layer to bottom to institute
It states decomposition image successively to repair, sample block restorative procedure first is based on to the decomposition image of top layer and is repaired, is repaired
Good top layer images repair result；Result up-sampling treatment is repaired to the top layer images, then decomposes image with secondary top layer and carries out
Merge, and the information in region to be filled in amalgamation result is repaired into time top layer as guidance information and decomposes image, repeats this mistake
Journey is repaired until all layers of regions to be filled and is finished.A kind of sample block image based on layering boot policy of the invention is repaired
Compound method, with applied widely, repairing effect is good and practical feature, the image that can be applicable in large area breakage are repaired
It is multiple, meet visual effect.
Detailed description of the invention
Fig. 1 is a kind of flow diagram of the sample block image repair method based on layering boot policy of the present invention.
Fig. 2 is a kind of picture breakdown schematic diagram of the sample block image repair method based on layering boot policy of the present invention.
Fig. 3 is a kind of region to be filled signal of the sample block image repair method based on layering boot policy of the present invention
Figure.
Fig. 4 is the schematic diagram after a kind of reparation of the sample block image repair method based on layering boot policy of the present invention.
Specific embodiment
Fig. 1 to Fig. 4 with reference to the accompanying drawing is to a kind of sample block image repair based on layering boot policy of the invention
Method is described in further detail.
A kind of sample block image repair method based on layering boot policy of the invention, please refers to Fig. 1-4, comprising:
Step 1: carrying out gaussian pyramid decomposition to original image to be repaired, the decomposition of multilayer different resolution is obtained
Image.Wherein, gaussian pyramid decomposition is one kind of multi-scale expression in image, and pyramid decomposition of the present invention takes sub-sampling,
That is K layers of gaussian pyramid are obtained with K+1 layers of Gaussian image by sub-sampling, cutoff frequency from upper one layer to
Next layer be gradually increased with the factor 2, so gaussian pyramid can cross over very big frequency range, as shown in FIG. 1, FIG. 1 is
Image after original image pyramid decomposition to be repaired.
Step 2: successively being repaired according to the sequence from top layer to bottom to the decomposition image, first to described in top layer points
Solution image is based on sample block restorative procedure and is repaired, and obtains the top layer images repaired and repairs result.Repairing based on sample block
Compound method repairs top layer images, and after pyramid decomposition image, the resolution ratio of top layer images is minimum in all images, breaks
The area for damaging region is minimum, so first repairing top layer images.Since the technology based on sample block can repair texture and knot simultaneously
Structure, therefore pyramidal top layer images are repaired using the method based on sample block, the image after reparation can keep texture and knot
The consistency of structure, more meets visual effect.
Step 3: repairing result up-sampling treatment to the top layer images, then image is decomposed with secondary top layer and is merged, and
Information in region to be filled in amalgamation result is repaired into time top layer as guidance information and decomposes image, repeats this process, until
It repairs and finishes in all layers of region to be filled.After pyramid decomposes image, from bottom to top layer, the resolution ratio of image gradually becomes
Low, the area of damaged area also accordingly becomes smaller, so successively repairing by sequence from top to bottom.After top layer images are repaired,
Repairing result can guide lower image further to repair, until image repair to be repaired is completed.
A kind of sample block image repair method based on layering boot policy of the invention, please refers to Fig. 1-4, retouches in front
If on the basis of the technical solution stated it may also is that filling forward positionOn have a point p, in the to be filled piece of Ψ centered on pp
In there are two region: Given information region ΨpaWith region Ψ to be filledpb, the restorative procedure based on sample block utilizes ΨpIn
ΨpaInformation, calculated according to formula (1) and in sample resource Φ search for Best exemplar patch
Wherein, Φ indicates the sample areas in image,It is and ΨpApart from the smallest sample block, d (Ψpa, Ψqa) table
Show ΨpaAnd ΨqaBetween color component and gradient component Euclidean distance；
Ψ is calculated further according to formula (2)pAnd ΨqBetween color component and gradient component Euclidean distance,
Wherein, Ip, IqIt is corresponding RGB vector,Indicate the gradient vector of image；
If top layer images GKS is obtained by up-sampling treatment after reparationK-1, SK-1With GK-1Merging obtains AK-1, at this moment AK-1In
Ω in information filled, using boot policy sample block repair image when, ΨpbIn information as guidance
Information, if to be filled piece is Ψp, Best exemplar patch is found at this timeBest exemplar patch is calculated using formula (3)
Wherein, ΨpaIt is Given information region, ΨpbIt is region to be filled, information therein is guidance information；
Distance is calculated using the colouring information of sample block by formula (4) again,
Wherein, Ip, IqIt is corresponding RGB vector.
When finding Best exemplar patch using formula (4), can use in all information, but can merely with colouring information
Best exemplar patch is found well, can also improve search speed, therefore formula (4) calculates distance without the gradient using sample block
Information.
It is above-mentioned that only several specific embodiments in the present invention are illustrated, but can not be as protection model of the invention
Enclose, it is all according to the present invention in the equivalent change or modification made of design spirit or equal proportion zoom in or out, should all
Think to fall into protection scope of the present invention.
Claims (1)
1. a kind of sample block image repair method based on layering boot policy, it is characterised in that: including
Step 1: carrying out gaussian pyramid decomposition to original image to be repaired, the decomposition image of multilayer different resolution is obtained；
Step 2: successively being repaired according to the sequence from top layer to bottom to the decomposition image, first to the exploded view of top layer
As being repaired based on sample block restorative procedure, obtains the top layer images repaired and repair result；
Step 3: repairing result up-sampling treatment to the top layer images, then image is decomposed with secondary top layer and is merged, and will close
And the information in result in region to be filled repairs time top layer as guidance information and decomposes image, this process is repeated, until all
It repairs and finishes in the region to be filled of layer；
It specifically includes:
If filling forward positionOn have a point p, in the to be filled piece of Ψ centered on ppIn there are two region: Given information region
ΨpaWith region Ψ to be filledpb, the restorative procedure based on sample block utilizes ΨpMiddle ΨpaInformation, according to formula (1) calculate exist
Best exemplar patch is searched in sample resource Φ
Wherein, Φ indicates the sample areas in image,It is and ΨpApart from the smallest sample block, d (Ψpa, Ψqa) indicate Ψpa
And ΨqaBetween color component and gradient component Euclidean distance；
Ψ is calculated further according to formula (2)pAnd ΨqBetween color component and gradient component Euclidean distance,
Wherein, Ip, IqIt is corresponding RGB vector,Indicate the gradient vector of image；
If top layer images GKS is obtained by up-sampling treatment after reparationK-1, SK-1With GK-1Merging obtains AK-1, at this moment AK-1In Ω
In information filled, using boot policy sample block repair image when, ΨpbIn information as guidance letter
Breath, if to be filled piece is Ψp, Best exemplar patch is found at this timeBest exemplar patch is calculated using formula (3)
Wherein, ΨpaIt is Given information region, ΨpbIt is region to be filled, information therein is guidance information；
Distance is calculated using the colouring information of sample block by formula (4) again,
Wherein, Ip, IqIt is corresponding RGB vector.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710082887.7A CN106875359B (en) | 2017-02-16 | 2017-02-16 | A kind of sample block image repair method based on layering boot policy |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710082887.7A CN106875359B (en) | 2017-02-16 | 2017-02-16 | A kind of sample block image repair method based on layering boot policy |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106875359A CN106875359A (en) | 2017-06-20 |
| CN106875359B true CN106875359B (en) | 2019-12-03 |
Family
ID=59166255
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710082887.7A Active CN106875359B (en) | 2017-02-16 | 2017-02-16 | A kind of sample block image repair method based on layering boot policy |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106875359B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109544465A (en) * | 2018-10-23 | 2019-03-29 | 天津大学 | Image damage block restorative procedure based on change of scale |
| CN110189278B (en) * | 2019-06-06 | 2020-03-03 | 上海大学 | Binocular scene image restoration method based on generation countermeasure network |
| CN110706167B (en) * | 2019-09-25 | 2022-06-10 | 中国人民解放军61646部队 | Fine completion processing method and device for remote sensing image to-be-repaired area |
| CN112258606B (en) * | 2020-10-21 | 2024-06-04 | 腾讯科技（深圳）有限公司 | Image processing method, device, equipment and readable storage medium |
| CN117974460B (en) * | 2024-03-29 | 2024-06-11 | 深圳中科精工科技有限公司 | Image enhancement method, system and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104182939A (en) * | 2014-08-18 | 2014-12-03 | 成都金盘电子科大多媒体技术有限公司 | Medical image detail enhancement method |
| CN106204502A (en) * | 2016-08-17 | 2016-12-07 | 重庆大学 | Based on mixing rank L0regularization fuzzy core method of estimation |
| CN106327449A (en) * | 2016-09-12 | 2017-01-11 | 厦门美图之家科技有限公司 | Image restoration method, image restoration application, and calculating equipment |
-
2017
- 2017-02-16 CN CN201710082887.7A patent/CN106875359B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104182939A (en) * | 2014-08-18 | 2014-12-03 | 成都金盘电子科大多媒体技术有限公司 | Medical image detail enhancement method |
| CN106204502A (en) * | 2016-08-17 | 2016-12-07 | 重庆大学 | Based on mixing rank L0regularization fuzzy core method of estimation |
| CN106327449A (en) * | 2016-09-12 | 2017-01-11 | 厦门美图之家科技有限公司 | Image restoration method, image restoration application, and calculating equipment |
Non-Patent Citations (4)
| Title |
|---|
| Exemplar based inpainting in a multi-scaled space;Baek-Sop Kim等;《Optik》;20151231;第3978-3981页 * |
| Exemplar-Based Image Inpainting Using Multiscale Graph Cuts;Yunqiang Liu等;《IEEE TRANSACTIONS ON IMAGE PROCESSING》;20130531;第22卷(第5期);第1699-1711页 * |
| HIERARCHICAL DIGITAL IMAGE INPAINTING USING WAVELETS;S.Padmavathi等;《Signal & Image Processing》;20120831;第3卷(第4期);第85-93页 * |
| 样本块搜索和优先权填充的弧形推进图像修复;刘华明等;《中国图象图形学报》;20161231;第21卷(第8期);第993-1003页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106875359A (en) | 2017-06-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106875359B (en) | A kind of sample block image repair method based on layering boot policy | |
| CN110135455B (en) | Image matching method, device and computer readable storage medium | |
| CN107103613B (en) | A kind of three-dimension gesture Attitude estimation method | |
| CN108898610A (en) | A kind of object contour extraction method based on mask-RCNN | |
| CN109903236B (en) | Face image restoration method and device based on VAE-GAN and similar block search | |
| CN107403424A (en) | A kind of car damage identification method based on image, device and electronic equipment | |
| CN106228528B (en) | A kind of multi-focus image fusing method based on decision diagram and rarefaction representation | |
| CN103824049A (en) | Cascaded neural network-based face key point detection method | |
| CN104408708B (en) | A kind of image well-marked target detection method based on global and local low-rank | |
| CN103971338B (en) | Variable-block image repair method based on saliency map | |
| CN107730507A (en) | A kind of lesion region automatic division method based on deep learning | |
| CN105913407B (en) | A method of poly focal power image co-registration is optimized based on differential chart | |
| CN104809698A (en) | Kinect depth image inpainting method based on improved trilateral filtering | |
| CN103886561B (en) | Criminisi image inpainting method based on mathematical morphology | |
| CN109034245A (en) | A kind of object detection method merged using characteristic pattern | |
| CN104537622B (en) | The method and system that raindrop influence is removed in single image | |
| Ting et al. | Image inpainting by global structure and texture propagation | |
| CN106530247A (en) | Multi-scale image restoring algorithm based on structure information | |
| CN108921942A (en) | The method and device of 2D transformation of ownership 3D is carried out to image | |
| CN110533713A (en) | Bridge Crack width high-precision measuring method and measuring device | |
| CN106570928B (en) | A kind of heavy illumination method based on image | |
| CN110751668B (en) | Image processing method, device, terminal, electronic equipment and readable storage medium | |
| CN107564013A (en) | Merge the scene cut modification method and system of local message | |
| Meeus et al. | Deep learning for paint loss detection with a multiscale, translation invariant network | |
| Song et al. | Building extraction from high resolution color imagery based on edge flow driven active contour and JSEG |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CP03 | Change of name, title or address | ||
| CP03 | Change of name, title or address |
Address after: 236000 Qinghe Road, Fuyang, Anhui Patentee after: Fuyang Normal University Country or region after: China Address before: 236037 Qinghe West Road, Fuyang, Anhui Province, No. 100 Patentee before: FUYANG NORMAL University Country or region before: China |
|
| TR01 | Transfer of patent right |
Effective date of registration: 20240603 Address after: Room 1905, Building 1, Science and Technology Entrepreneurship City, No. 17 Huaguoshan Avenue, Haizhou District, Lianyungang City, Jiangsu Province, 222000 Patentee after: Rongcheng Digital Technology (Lianyungang) Co.,Ltd. Country or region after: China Address before: 236000 Qinghe Road, Fuyang, Anhui Patentee before: Fuyang Normal University Country or region before: China |