CN102567955B - Method and system for inpainting images - Google Patents
Method and system for inpainting images Download PDFInfo
- Publication number
- CN102567955B CN102567955B CN201010616564.XA CN201010616564A CN102567955B CN 102567955 B CN102567955 B CN 102567955B CN 201010616564 A CN201010616564 A CN 201010616564A CN 102567955 B CN102567955 B CN 102567955B
- Authority
- CN
- China
- Prior art keywords
- point
- repaired
- image
- neighborhood
- setting
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Landscapes
- Image Processing (AREA)
Abstract
The invention provides a method and a system for inpainting images, which are used for solving the problems that an image inpainting effect in the prior art is poorer. The method comprises the following steps of: determining the outline of a to-be-inpainted region; respectively inpainting unknown pixel points of all outline point setting neighbourhoods of the to-be-inpainted region of the image to obtain an inpainted image, wherein the obtained inpainted image is used as an updated image; repeating the two steps, i.e. inpainting the to-be-inpainted region of the image from edges inwards in sequence layer by layer until the to-be-inpainted region is completely filled.
Description
Technical field
The present invention relates to a kind of method and system of image repair.
Background technology
Along with the development of computing machine and digital technology, digital picture is widely used by people with advantages such as its convenient preservation and processing.But digital picture is in the loss of data in storage medium, in compression, transmitting procedure, all likely cause the damage of digital picture because of the interference of extraneous noise or the loss of self information.The old photo of damage that has scratch or a spot spot also needs it to repair being scanned into after digital picture.In addition, due to special requirement, need to remove earnest body or word in image, the white space staying after removing need to be filled up.Therefore, people are urgent expect to obtain one simply, automatically, image edit method fast, complete easily image repair.
At present, both at home and abroad for the solution thinking of image repair problem mainly comprise image repair based on variation nonlinear partial differential equation PDE (Partial Differential Equations), based on the synthetic image repair of texture.Wherein the image repair based on variation nonlinear partial differential equation is repaired cavity in image along isophote direction by the mode spreading gradually by area to be repaired peripheral information.Based on the synthetic image repair method of texture using image not drop-out as derivation graph the training set as drop-out, in training set, find the sample of coupling and insert Image blank district, realize effective reparation of image.
Repair unstable result based on the image repair method of variation nonlinear partial differential equation, can not effectively repair large lost regions, have obvious distortion at the center of repairing result, repairing effect is not good enough, and the method need to iterate, and efficiency is not high.Easily cause the mistake breeding of picture structure based on the synthetic image repair method of texture, repairing effect is not ideal enough, and the calculating of itself is more consuming time, causes the efficiency of image repair lower.
There is the problem that repairing effect is poor and remediation efficiency is lower in the method for image repair of the prior art, for this problem, not yet proposes at present effective solution.
Summary of the invention
Fundamental purpose of the present invention is to provide a kind of method and system of image repair, and to solve the poor problem of image repair effect in prior art, another object of the present invention is to solve the lower problem of image repair efficiency in prior art.
To achieve these goals, according to an aspect of the present invention, provide a kind of method of image repair.
The method of image repair of the present invention comprises: the profile of determining image area to be repaired; The unknown pixel point of all point setting neighborhoods to area to be repaired described in image is repaired respectively, and the image that reparation is obtained is as after image; Repeat above-mentioned two steps, to image area to be repaired inside from edge carry out successively layering reparation, until that all fill area to be repaired is complete.
The described unknown pixel point of respectively all point of described area to be repaired being set to neighborhoods is repaired and is comprised: A, according to the image information of the known pixels point in the point setting neighborhood of image area to be repaired, is divided into level and smooth point and structure outline point by point; B, level and smooth point to described image area to be repaired and the setting neighborhood unknown pixel point of structure outline point are repaired.
Further, the described point according to image area to be repaired is set the image composition of the known pixels point of neighborhood, point is divided into level and smooth point and structure outline point comprises: determine the setting neighborhood of putting centered by point p; The gradient of calculating known pixels point in described setting neighborhood, the weighted gradient amplitude of adding up each gradient direction, gets the weighted gradient amplitude of maximal value as point p, and described gradient direction is the direction obtaining through normalization; As described in weighted gradient amplitude be less than setting threshold, this point is level and smooth point, on the contrary this point is structure outline point.
Further, the weighted gradient amplitude of described each gradient direction of statistics is according to by setting each known pixels point and the weights that the distance of point p determines in neighborhood, is weighted the numerical value that summation obtains, and the less weights of its middle distance are larger.
Further, the unknown pixel point that described structure outline point is set to neighborhood is repaired and is comprised: according to the weighted gradient amplitude of the setting neighborhood known pixels point of structure outline point, by by force to the weak unknown pixel point of repairing successively each structure outline point.
Further, the unknown pixel point of described level and smooth point being set to neighborhood is repaired and is comprised: adopt the mode of mirror image to fill the unknown pixel point of the setting neighborhood of described level and smooth point.
Further, the unknown pixel point of the setting neighborhood to described structure outline point is repaired and is comprised: adopt the synthetic method of texture to fill the unknown pixel point of the setting neighborhood of described structure outline point.
Further, in described step B, set the overlapping part of the unknown pixel point of neighborhood and the unknown pixel point of structure outline point setting neighborhood for level and smooth point, the result that adopts structure outline point to fill.
A kind of system of image repair is provided according to a further aspect in the invention.
The system of image repair of the present invention comprises: profile determination module, for determining the profile of image area to be repaired; Repair operational module, repair for the unknown pixel point of respectively all point of area to be repaired described in image being set to neighborhoods, the image that reparation is obtained is as after image; Repair control module, for control described profile determination module and described reparation operational module to image area to be repaired inside from edge carry out successively layering reparation, until that all fill area to be repaired is complete.
Further, described reparation operational module also for: A, set the image information of the known pixels point in neighborhood according to the point of image area to be repaired, point is divided into level and smooth point and structure outline point; B, level and smooth point to described image area to be repaired and the setting neighborhood unknown pixel point of structure outline point are repaired.
Further, described reparation operational module also for: determine the setting neighborhood of putting centered by point p; The gradient of calculating known pixels point in described setting neighborhood, the weighted gradient amplitude of adding up each gradient direction, gets the weighted gradient amplitude of maximal value as point p, and described gradient direction is the direction obtaining through normalization; As described in weighted gradient amplitude be less than setting threshold, this point is level and smooth point, on the contrary this point is structure outline point.
Further, described reparation operational module also, for according to by setting each known pixels point and the weights that the distance of point p determines in neighborhood, is weighted the numerical value that summation obtains, and the less weights of its middle distance are larger.
Further, described reparation operational module is also for the weighted gradient amplitude of the setting neighborhood known pixels point according to structure outline point, by by force to the weak unknown pixel point of repairing successively each structure outline point.
Further, described reparation operational module is also for adopting the mode of mirror image to fill the unknown pixel point of the setting neighborhood of described level and smooth point.
Further, described reparation operational module is also for adopting the synthetic method of texture to fill the unknown pixel point of the setting neighborhood of described structure outline point.
Further, described reparation operational module is also for setting the overlapping part of the unknown pixel point of neighborhood and the unknown pixel point of structure outline point setting neighborhood, the result that adopts structure outline point to fill for level and smooth point.
Can find out from above explanation, layering processing is carried out in the area to be repaired for image in the technical scheme of the present embodiment, contributes to like this to reduce the construction error breeding of sample repair process, thereby has improved the effect of image repair.And in the embodiment of the present invention, repair in different ways for level and smooth point and structure outline point, contribute to improve the efficiency of image repair.
Brief description of the drawings
Figure of description is used to provide a further understanding of the present invention, forms the application's a part, and schematic description and description of the present invention is used for explaining the present invention, does not form inappropriate limitation of the present invention.In the accompanying drawings:
Fig. 1 is the key step process flow diagram according to the method for the image repair of the embodiment of the present invention;
Fig. 2 is the schematic diagram that image is carried out to layering reparation according to the embodiment of the present invention;
Fig. 3 is according to the schematic diagram of a kind of idiographic flow of the image repair method of the embodiment of the present invention; And
Fig. 4 is the main modular schematic diagram according to the system of the image repair of the embodiment of the present invention.
Embodiment
It should be noted that, in the situation that not conflicting, the feature in embodiment and embodiment in the application can combine mutually.Describe below with reference to the accompanying drawings and in conjunction with the embodiments the present invention in detail.
Fig. 1 is the key step process flow diagram according to the method for the image repair of the embodiment of the present invention, and as shown in Figure 1, the method mainly comprises the steps:
Step S11: the profile of determining image area to be repaired;
Step S13: the unknown pixel point of all point setting neighborhoods to image area to be repaired is repaired respectively, and the image that reparation is obtained is as after image;
Step S15: judge whether area to be repaired all fills complete; If so, process ends, otherwise return to step S11.
Can find out from above-mentioned flow process, layering processing is carried out in the area to be repaired for image in the technical scheme of the present embodiment, contributes to like this to reduce the construction error breeding of sample repair process, thereby has improved the effect of image repair.The thought that layering is repaired can be as seen from Figure 2.Fig. 2 is the schematic diagram that image is carried out to layering reparation according to the embodiment of the present invention.
As shown in Figure 2, the failure area of square frame 21 presentation videos 20, this failure area is initial area to be repaired, while starting to repair image, in the time of first pass execution step S11 and step S13, reparation be the region between square frame 21 and square frame 22, while again performing step S11 and step S13, what repair is the region between square frame 22 and square frame 23, and the rest may be inferred, until the whole reparations of the failure area shown in square frame 21 are complete.
In the present embodiment, in order to improve the efficiency of image repair, in the time of the reparation operation of carrying out area to be repaired, the level and smooth point to area to be repaired and structure outline point are processed respectively, specifically can be divided into following two steps:
Steps A, set the image information of the known pixels point in neighborhood according to the point of image area to be repaired, point is divided into level and smooth point and structure outline point; Step B, level and smooth point to image area to be repaired and the setting neighborhood unknown pixel point of structure outline point are repaired.Below again steps A and step B are described further.
In steps A, can first determine the setting neighborhood of putting centered by point p; Then calculate the gradient of setting known pixels point in neighborhood, the weighted gradient amplitude of adding up each gradient direction, gets the weighted gradient amplitude of maximal value as point p; As weighted gradient amplitude is less than setting threshold, this point is level and smooth point, otherwise this point is structure outline point.The weighted gradient amplitude of adding up each gradient direction here can be according to by setting each known pixels point and the weights that the distance of point p determines in neighborhood, is weighted the numerical value that summation obtains, and the less weights of its middle distance are larger.
When structure outline point is set the unknown pixel point of neighborhood and repaired, can be specifically according to the weighted gradient amplitude of the setting neighborhood known pixels point of structure outline point, by by force to the weak unknown pixel point of repairing successively each structure outline point.
On concrete repair mode, for level and smooth point and structure outline point in different ways, to improve remediation efficiency.In the present embodiment, when the unknown pixel point of level and smooth point setting neighborhood is repaired, adopt the mode of mirror image to fill the unknown pixel point of the setting neighborhood of level and smooth point; When the unknown pixel point of the setting neighborhood to structure outline point is repaired, adopt the unknown pixel point of the setting neighborhood of the synthetic method interstitital texture point of texture.In above-mentioned steps B, in image, there is sometimes the overlapping part of the unknown pixel point of level and smooth point setting neighborhood and the unknown pixel point of structure outline point setting neighborhood.For this part image, the result that adopts structure outline point to fill in the present embodiment, to ensure the relative precision of image repair.
Fig. 3 is according to the schematic diagram of a kind of idiographic flow of the image repair method of the embodiment of the present invention, below this flow process is made an explanation.
In the time starting to process image, color of image pattern is converted to Lab pattern.This is to consider that the expression of the image in Lab space more meets the visual characteristic of human eye, so carry out the coupling of image similarity piece in Lab space, improves the correct matching degree of similar with this, reduces the wrong structure not being inconsistent with human eye and extends.
Next determine the profile of area to be repaired, then analyze the iconic element of the known pixels of each vertex neighborhood on the profile of area to be repaired, point is divided into level and smooth and structure two classes.
In the time that point is classified, specifically can adopt with the following method: to 1 p on profile, centered by it, get neighborhood piece P.Calculate the gradient of known pixels point in this piece, add up the direction histogram of these gradients, obtain according to histogram of gradients direction and the corresponding intensity of this direction that the neighborhood P of a p comprises structure, described direction refers to that gradient direction is the direction that normalization obtains, as 360 degree being normalized to 8 main directions, 0~45 degree is thought and is belonged to same direction.Wherein, the value of histogram i post be direction be i gradient amplitude weighting and, weights according to each known pixels point q in neighborhood piece with some p distance determine, distance less weights larger, apart from larger weights less.On using this weighted gradient amplitude as profile to be repaired, each point comprises the how many tolerance of structure.If this value is got over close to zero, just to think and change the time as level and smooth point, its neighborhood is likely also level and smooth.Otherwise, if this value is very large, can think that current point is structure outline point, its neighborhood may comprise abundant structure.
Adopt the mode of mirror image that known pixels is repaired to area to be repaired for level and smooth point, until repaired all level and smooth point.For structure outline point, the how many arrangement reparations order that comprises constituent according to its neighborhood known pixels.
In the time that structure outline point is repaired, can adopt the synthetic method of texture.Can be specifically to take out the optimum neighborhood of a point piece of filling, taking the known pixels point in this neighborhood piece as reference, search the most similar image block with it in image known region; The pixel value of the corresponding pixel points in the similar image piece that the unknown pixel in optimum filling point neighborhood is put to search is filled, fill later unknown pixel point while once filling on carrying out as known pixels point.For similar point: fill pixel later and think known; But the foundation when reparation result of smooth region is not repaired as structural region in the present embodiment.For example, if the pixel that has comprised the smooth region being filled in a structure outline neighborhood of a point, the foundation of coupling when these points are not repaired as this structure outline point so, and as unknown pixel, this is also for fear of the filling result of smooth region, the mistake of structural region to be covered.That is to say, in the process of above-mentioned reparation, if overlapping if the reparation result of the reparation result of system point and level and smooth point has, cover with the reparation result of system point, emphasize the reparation of system point.So far the image repair of one deck is complete.Then upgrade the profile of area to be repaired, start the reparation of lower one deck; Repeat above-mentioned steps until area to be repaired is all filled complete.
Fig. 4 is the main modular schematic diagram according to the system of the image repair of the embodiment of the present invention.As shown in Figure 4, the system 40 of image repair mainly comprises as lower module: profile determination module, for determining the profile of image area to be repaired; Repair operational module, repair for the unknown pixel point of respectively all point of image area to be repaired being set to neighborhoods, the image that reparation is obtained is as after image; Repair control module, for control profile determination module and repair operational module to image area to be repaired inside from edge carry out successively layering reparation, until that all fill area to be repaired is complete.
Repair the image information that operational module also can be used for repairing as follows operation: A, sets the known pixels point in neighborhood according to the point of image area to be repaired, point is divided into level and smooth point and structure outline point; B, first the setting neighborhood unknown pixel point of the level and smooth point to image area to be repaired is repaired, and again, the unknown pixel point of the setting neighborhood of the structure outline point to image area to be repaired is repaired.
Repair operational module and also can be used for determining the setting neighborhood of putting centered by point p; And calculate the gradient of setting known pixels point in neighborhood, the weighted gradient amplitude of adding up each gradient direction, gets the weighted gradient amplitude of maximal value as point p; As weighted gradient amplitude is less than setting threshold, this point is level and smooth point, otherwise this point is structure outline point.
Repair operational module and also can be used for, according to by setting each known pixels point and the weights that the distance of point p determines in neighborhood, being weighted the numerical value that summation obtains, the less weights of its middle distance are larger.
Repair operational module and also can be used for according to the weighted gradient amplitude of the setting neighborhood known pixels point of structure outline point, by by force to the weak unknown pixel point of repairing successively each structure outline point.
Repair operational module and also can be used for adopting the mode of mirror image to fill the unknown pixel point of the setting neighborhood of level and smooth point, or adopt the unknown pixel point of the setting neighborhood of the synthetic method interstitital texture point of texture.
Repair operational module and also can be used for the overlapping part for the unknown pixel point of level and smooth point setting neighborhood and the unknown pixel point of structure outline point setting neighborhood, the result that adopts structure outline point to fill.
Can find out from above explanation, layering processing is carried out in the area to be repaired for image in the technical scheme of the present embodiment, contributes to like this to reduce the construction error breeding of sample repair process, thereby has improved the effect of image repair.And in the embodiment of the present invention, repair in different ways for level and smooth point and structure outline point, contribute to improve the efficiency of image repair.
Obviously, those skilled in the art should be understood that, above-mentioned of the present invention each module or each step can realize with general calculation element, they can concentrate on single calculation element, or be distributed on the network that multiple calculation elements form, alternatively, they can be realized with the executable program code of calculation element, thereby, they can be stored in memory storage and be carried out by calculation element, or they are made into respectively to each integrated circuit modules, or the multiple modules in them or step are made into single integrated circuit module to be realized.Like this, the present invention is not restricted to any specific hardware and software combination.
The foregoing is only the preferred embodiments of the present invention, be not limited to the present invention, for a person skilled in the art, the present invention can have various modifications and variations.Within the spirit and principles in the present invention all, any amendment of doing, be equal to replacement, improvement etc., within all should being included in protection scope of the present invention.
Claims (12)
1. a method for image repair, is characterized in that, comprising:
Determine the profile of image area to be repaired;
The unknown pixel point of all point setting neighborhoods to area to be repaired described in image is repaired respectively, and the image that reparation is obtained is as after image;
Repeat above-mentioned two steps, to image area to be repaired inside from edge carry out successively layering reparation, until that all fill area to be repaired is complete;
Wherein, the described unknown pixel point of respectively all point of described area to be repaired being set to neighborhoods is repaired and is comprised: A, according to the image information of the known pixels point in the point setting neighborhood of image area to be repaired, is divided into level and smooth point and structure outline point by point; B, level and smooth point to described image area to be repaired and the setting neighborhood unknown pixel point of structure outline point are repaired;
Wherein, the described point according to image area to be repaired is set the image composition of the known pixels point of neighborhood, point is divided into level and smooth point and structure outline point comprises: determine the setting neighborhood of putting centered by point p; The gradient of calculating known pixels point in described setting neighborhood, the weighted gradient amplitude of adding up each gradient direction, gets the weighted gradient amplitude of maximal value as point p, and described gradient direction is the direction obtaining through normalization; As described in weighted gradient amplitude be less than setting threshold, this point is level and smooth point, on the contrary this point is structure outline point.
2. method according to claim 1, it is characterized in that, the weighted gradient amplitude of described each gradient direction of statistics is according to by setting each known pixels point and the weights that the distance of point p determines in neighborhood, is weighted the numerical value that summation obtains, and the less weights of its middle distance are larger.
3. method according to claim 1, it is characterized in that, the unknown pixel point that described structure outline point is set to neighborhood is repaired and is comprised: according to the weighted gradient amplitude of the setting neighborhood known pixels point of structure outline point, by by force to the weak unknown pixel point of repairing successively each structure outline point.
4. method according to claim 1, is characterized in that, the unknown pixel point that described level and smooth point is set to neighborhood is repaired and comprised: adopt the mode of mirror image to fill the unknown pixel point of the setting neighborhood of described level and smooth point.
5. according to the method described in claim 1 or 3, it is characterized in that, the unknown pixel point of the setting neighborhood to described structure outline point is repaired and is comprised: adopt the synthetic method of texture to fill the unknown pixel point of the setting neighborhood of described structure outline point.
6. according to the method described in claim 1 or 3, it is characterized in that, in described step B, set the overlapping part of the unknown pixel point of neighborhood and the unknown pixel point of structure outline point setting neighborhood for level and smooth point, the result that adopts structure outline point to fill.
7. a system for image repair, is characterized in that, comprising:
Profile determination module, for determining the profile of image area to be repaired;
Repair operational module, repair for the unknown pixel point of respectively all point of area to be repaired described in image being set to neighborhoods, the image that reparation is obtained is as after image;
Repair control module, for control described profile determination module and described reparation operational module to image area to be repaired inside from edge carry out successively layering reparation, until that all fill area to be repaired is complete;
Wherein, described reparation operational module also for: A, set the image information of the known pixels point in neighborhood according to the point of image area to be repaired, point is divided into level and smooth point and structure outline point; B, level and smooth point to described image area to be repaired and the setting neighborhood unknown pixel point of structure outline point are repaired;
Wherein, described reparation operational module also for: determine the setting neighborhood of putting centered by point p; The gradient of calculating known pixels point in described setting neighborhood, the weighted gradient amplitude of adding up each gradient direction, gets the weighted gradient amplitude of maximal value as point p, and described gradient direction is the direction obtaining through normalization; As described in weighted gradient amplitude be less than setting threshold, this point is level and smooth point, on the contrary this point is structure outline point.
8. system according to claim 7, is characterized in that, described reparation operational module is also for according to by setting in neighborhood
Each known pixels point and the weights that the distance of point p determines, be weighted the numerical value that summation obtains, and the little weights of its middle distance are larger.
9. system according to claim 7, is characterized in that, described reparation operational module is also for the weighted gradient amplitude of the setting neighborhood known pixels point according to structure outline point, by by force to the weak unknown pixel point of repairing successively each structure outline point.
10. system according to claim 7, is characterized in that, described reparation operational module is also for adopting the mode of mirror image to fill the unknown pixel point of the setting neighborhood of described level and smooth point.
11. according to the system described in claim 7 or 9, it is characterized in that, described reparation operational module is also for adopting the synthetic method of texture to fill the unknown pixel point of the setting neighborhood of described structure outline point.
12. according to the system described in claim 7 or 9, it is characterized in that, described reparation operational module is also for setting the overlapping part of the unknown pixel point of neighborhood and the unknown pixel point of structure outline point setting neighborhood, the result that adopts structure outline point to fill for level and smooth point.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201010616564.XA CN102567955B (en) | 2010-12-30 | 2010-12-30 | Method and system for inpainting images |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201010616564.XA CN102567955B (en) | 2010-12-30 | 2010-12-30 | Method and system for inpainting images |
Publications (2)
Publication Number | Publication Date |
---|---|
CN102567955A CN102567955A (en) | 2012-07-11 |
CN102567955B true CN102567955B (en) | 2014-09-10 |
Family
ID=46413309
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201010616564.XA Expired - Fee Related CN102567955B (en) | 2010-12-30 | 2010-12-30 | Method and system for inpainting images |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN102567955B (en) |
Families Citing this family (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102968764B (en) * | 2012-10-26 | 2015-04-29 | 北京航空航天大学 | Chinese character image inpainting method based on strokes |
CN105761213B (en) * | 2014-12-16 | 2019-02-26 | 北京大学 | Image mending method and image mending device |
CN105427233A (en) * | 2015-12-29 | 2016-03-23 | 小米科技有限责任公司 | Method and device for removing watermark |
CN106228519B (en) * | 2016-07-18 | 2019-08-30 | Oppo广东移动通信有限公司 | A kind of image repair method and terminal |
CN107358581A (en) * | 2017-06-19 | 2017-11-17 | 东南大学 | Rapid image restorative procedure |
CN109142834B (en) * | 2017-06-19 | 2022-01-28 | 北京普源精电科技有限公司 | Amplitude weight obtaining method and device for track drawing and oscilloscope |
CN107833196A (en) * | 2017-12-19 | 2018-03-23 | 蒙城县望槐信息科技有限责任公司 | A kind of image deflects point circle restorative procedure |
CN108022224A (en) * | 2017-12-19 | 2018-05-11 | 蒙城县望槐信息科技有限责任公司 | A kind of image repair system |
CN108765295B (en) * | 2018-06-12 | 2019-11-26 | 腾讯科技(深圳)有限公司 | Image processing method, image processing apparatus and storage medium |
KR102492369B1 (en) | 2018-08-21 | 2023-01-27 | 후아웨이 테크놀러지 컴퍼니 리미티드 | Binarization and normalization-based inpainting for text removal |
CN110310235B (en) * | 2019-05-21 | 2021-07-27 | 北京至真互联网技术有限公司 | Fundus image processing method, device and equipment and storage medium |
CN110599511B (en) * | 2019-08-20 | 2023-09-01 | 广东奥珀智慧家居股份有限公司 | Image filtering method, device and storage medium for reserving image edges |
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 |
CN113643437A (en) * | 2021-08-24 | 2021-11-12 | 凌云光技术股份有限公司 | Method and device for correcting depth image protrusion interference noise |
Family Cites Families (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN100458846C (en) * | 2005-07-14 | 2009-02-04 | 北京航空航天大学 | A method of image restoration |
CN101635047A (en) * | 2009-03-25 | 2010-01-27 | 湖南大学 | Texture synthesis and image repair method based on wavelet transformation |
CN101777178B (en) * | 2010-01-28 | 2011-08-31 | 南京大学 | Image restoring method |
-
2010
- 2010-12-30 CN CN201010616564.XA patent/CN102567955B/en not_active Expired - Fee Related
Also Published As
Publication number | Publication date |
---|---|
CN102567955A (en) | 2012-07-11 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN102567955B (en) | Method and system for inpainting images | |
CN104995911B (en) | Image processing device, image capture device, filter generating device, image restoration method, and program | |
US11790499B2 (en) | Certificate image extraction method and terminal device | |
US20100098331A1 (en) | System and method for segmenting foreground and background in a video | |
US9704227B2 (en) | Method and apparatus for image enhancement | |
CN109711407B (en) | License plate recognition method and related device | |
CN104517110A (en) | Binarization method and system of two-dimensional code image | |
CN105408936A (en) | System and method of correcting image artifacts | |
US20130002810A1 (en) | Outlier detection for colour mapping | |
CN104463161A (en) | Color document image segmentation and binarization using automatic inpainting | |
US8818056B2 (en) | Method and device for acquiring latent fingerprint orientation | |
Kim et al. | Multiple level feature-based universal blind image quality assessment model | |
CN111353956B (en) | Image restoration method and device, computer equipment and storage medium | |
CN106875408A (en) | Method, device and terminal device for sectional drawing | |
CN103020924A (en) | Low-illumination monitored image enhancement method based on similar scenes | |
CN105550981A (en) | Image registration and splicing method on the basis of Lucas-Kanade algorithm | |
CN116681636A (en) | Light infrared and visible light image fusion method based on convolutional neural network | |
CN110838088B (en) | Multi-frame noise reduction method and device based on deep learning and terminal equipment | |
CN113888509A (en) | Method, device and equipment for evaluating image definition and storage medium | |
CN104182976A (en) | Field moving object fining extraction method | |
CN113808033A (en) | Image document correction method, system, terminal and medium | |
CN113674303A (en) | Image processing method, image processing device, electronic equipment and storage medium | |
CN115170489A (en) | Method and device for identifying and measuring steel bars | |
JP3897306B2 (en) | Method for supporting extraction of change region between geographic images and program capable of supporting extraction of change region between geographic images | |
CN115242979A (en) | Focusing method and system applied to blood shooting, intelligent equipment and storage medium |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20140910 Termination date: 20141230 |
|
EXPY | Termination of patent right or utility model |