CN107194897A - A kind of precedence algorithm and image repair method based on architectural difference Yu marginal texture coefficient to be repaired - Google Patents
A kind of precedence algorithm and image repair method based on architectural difference Yu marginal texture coefficient to be repaired Download PDFInfo
- Publication number
- CN107194897A CN107194897A CN201710438763.8A CN201710438763A CN107194897A CN 107194897 A CN107194897 A CN 107194897A CN 201710438763 A CN201710438763 A CN 201710438763A CN 107194897 A CN107194897 A CN 107194897A
- Authority
- CN
- China
- Prior art keywords
- mrow
- repaired
- msub
- area
- point
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000004422 calculation algorithm Methods 0.000 title claims abstract description 42
- 230000008439 repair process Effects 0.000 title claims abstract description 32
- 238000000034 method Methods 0.000 title claims abstract description 16
- 239000011159 matrix material Substances 0.000 claims abstract description 23
- 239000000284 extract Substances 0.000 claims abstract description 8
- 238000000605 extraction Methods 0.000 claims abstract description 7
- 238000004364 calculation method Methods 0.000 claims description 11
- 230000010354 integration Effects 0.000 claims description 2
- 230000000694 effects Effects 0.000 description 6
- 230000007547 defect Effects 0.000 description 4
- 230000000007 visual effect Effects 0.000 description 4
- 241000208340 Araliaceae Species 0.000 description 3
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 3
- 235000003140 Panax quinquefolius Nutrition 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 235000008434 ginseng Nutrition 0.000 description 3
- 238000012152 algorithmic method Methods 0.000 description 2
- 238000006073 displacement reaction Methods 0.000 description 2
- 238000012163 sequencing technique Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000001427 coherent effect Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 210000001061 forehead Anatomy 0.000 description 1
- 210000004209 hair Anatomy 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- G06T5/77—
Abstract
The invention discloses a kind of precedence algorithm and image repair method based on architectural difference Yu marginal texture coefficient to be repaired, algorithm steps include:Damaged image is read, and marks area to be repaired;Extract all known pixels point information in target area;According to the known pixels of extraction point information architecture architectural difference matrix;A certain size the window area centered on any point p to be repaired on the border of area to be repaired is extracted in target area, complex point p to be repaired marginal texture coefficient T (p), standard deviation criteria S is calculatedpAnd confidence level parameter C (p) (p);According to formula P (p)=T (p) C (p) Sp(p) the priority valve P (p) of this complex point to be repaired is calculated;All complex points to be repaired on the border of area to be repaired are traveled through, the window area centered on the maximum complex point to be repaired of priority valve are repaired, and update area to be repaired border;Repeat the above steps, until area to be repaired repairs complete.Difference size of the invention according to each target area, the priority valve stablized greatly reduces the calculating time of priority.
Description
Technical field
The present invention relates to technical field of image processing, one kind is in particulard relate to based on architectural difference and marginal texture to be repaired
The precedence algorithm and image repair method of coefficient.
Background technology
Image repair is an ancient art, originating from European the Renaissance, is often referred to recover to lose in image
Or damaged pixel or region.People are very sensitive to discontinuous border, and lost regions generally comprise structure, texture
With it is smooth, structure is coherent closer to real visual effect, therefore researcher separates this three layers, and it is wherein to recover structure sheaf
Important part.Influence of the selection of reparation order to repairing result is very big, and the general block with structural type is should be by
Repair first.Criminisi utilizes the isophote direction of local boundary and the fill order of unit normal direction definition block,
After the sample block preferentially filled, best matching blocks are found in known region using global search mode, and by optimal
Match block fills sample block to be repaired.
Also certain methods choose image block centered on utilizing area to be repaired boundary point, then with selected block and neighborhood block
Difference size calculate similarity weight, with all known neighborhood blocks similarity weight build structure degree of rarefication, use structure
Degree of rarefication calculates fill order.Repair block build be the rarefaction representation using multiple match blocks as filling information, improve reparation
The quality of image.In addition, also similar block can be calculated in the known region of image, and block compensation is calculated, most image has
Similar compensation, the peak value for compensating distribution is related to main compensation.According to main compensation, input picture, which is formd, much has displacement
Version image.These images for having displacement are merged by optimizing global MRF energy functions.These methods are to repairing image production
Good visual effect and connectedness are given birth to, particularly, the defect area being repaired has uniform texture and a small number of structures.
The image repair problem removed for object, some scholars propose the image being combined based on sample block and depth
Repair algorithm, find the order that purpose image repairs object with the depth map of image, and object various visual angles figure, then go to fill out
Perforations adding hole.In order to reduce the size of database, it is proposed that main perspective extracts the method for (Keyview Extraction), and uses
A kind of geometry and flash ranging degree algorithm carries out object retrieval.Also the method based on example, is improved to priority feature,
The priority of linear structure is considered in pri function, and makes algorithm linear structure earlier, with solve completion order and
Linear structure propagation problem.In addition, in order to detect maximally effective structure, its structure tensor and image have been used in priority
The combination of gradient.
Although above-mentioned algorithm can realize image repair function to a certain extent, algorithm can not overcome in the presence of some
The problem of, such as fill order is not sufficiently stable, matching criterior is not reasonable, time-consuming longer.
The content of the invention
In view of the deficiency that prior art is present, a kind of using between small image block the invention aims to provide
The precedence algorithm of difference size discrimination complex point region to be repaired.
To achieve these goals, technical solution of the present invention is as follows:
A kind of precedence algorithm based on architectural difference Yu marginal texture coefficient to be repaired, it is characterised in that its step bag
Include:
S1 reads damaged image, and marks area to be repaired;
S2 extractions are included including the border of area to be repaired, and the certain target area of size, and extract institute in target area
There is known pixels point information;
S3 is according to the known pixels point information architecture architectural difference matrix of extraction;
S4 extracts a certain size the window centered on any point p to be repaired on the border of area to be repaired in target area
Mouth region domain, calculates complex point p to be repaired marginal texture coefficient T (p);
S5 calculates standard deviation criteria Ss of the complex point p to be repaired in the window area centered on itp(p) and confidence level ginseng
Number C (p);
It is to be repaired that S6 calculates this according to marginal texture coefficient, complex point standard deviation criteria to be repaired and pixel confidence parameter to be repaired
The priority valve P (p) of point, calculation formula is P (p)=T (p) C (p) Sp(p);
S7 repeat step S5-S6, until all complex points to be repaired on traversal area to be repaired border, obtain corresponding priority
Set;
S8 repairs the window area centered on the maximum complex point to be repaired of priority valve, and update area to be repaired border and
Architectural difference matrix;
S9 repeat step S4-S8, until area to be repaired repairs complete.
It is another object of the present invention to provide a kind of image repair method based on above-mentioned precedence algorithm, its feature exists
In comprising the following steps:
Step 1, reading restored image to be repaired, and determine area to be repaired boundary point;
Step 2, the calculating each corresponding priority valve of complex point to be repaired in area to be repaired border;
Step 3, the complex point to be repaired for searching priority valve maximum, and repair the target area figure centered on this complex point to be repaired
As block;
Step 4, refreshing architectural difference matrix, judge whether the region do not repaired also, if so, then performing step 2- steps
Rapid 3, until image repair is complete.
Compared with prior art, beneficial effects of the present invention:
The present invention is according to the position characteristics of the difference size discrimination complex point to be repaired of each target area, the priority stablized
Value;The difference between target area is deposited in the way of standard deviation in architectural difference matrix simultaneously, due to adjacent complex point to be repaired
Neighborhood exist and significantly overlap, repeat and calculate priority using the information in architectural difference matrix, so as to substantially reduce
Calculating time of priority.
Brief description of the drawings
In order to illustrate more clearly about the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing
There is the accompanying drawing used required in technology description to be briefly described, it should be apparent that, drawings in the following description are this hairs
Some bright embodiments, for those of ordinary skill in the art, without having to pay creative labor, can be with
Other accompanying drawings are obtained according to these accompanying drawings.
Fig. 1 is precedence algorithm flow chart of the present invention;
Fig. 2 a are image texture example region figure;
Fig. 2 b are picture structure example region figure;
Fig. 3 is image repair method flow chart of the present invention;
Fig. 4 is the present invention and contrast algorithm repairing effect comparison diagram.
Embodiment
To make the purpose, technical scheme and advantage of the embodiment of the present invention clearer, below in conjunction with the embodiment of the present invention
In accompanying drawing, the technical scheme in the embodiment of the present invention is clearly and completely described, it is clear that described embodiment is
A part of embodiment of the present invention, rather than whole embodiments.Based on the embodiment in the present invention, those of ordinary skill in the art
The every other embodiment obtained under the premise of creative work is not made, belongs to the scope of protection of the invention.
The present invention devises a kind of difference size discrimination complex point region to be repaired using between small image block
Precedence algorithm and image repair method, below in conjunction with the accompanying drawings and specific embodiment further illustrate the present invention technical side
Case:
As shown in figure 1, the step of precedence algorithm based on architectural difference and marginal texture coefficient to be repaired includes:
S1 reads damaged image, and marks area to be repaired.
S2 extractions are included including the border of area to be repaired, and the certain target area of size, and extract institute in target area
There is known pixels point information.
S3 includes according to the known pixels point information architecture architectural difference matrix of extraction, step:
Breakage image is accordingly divided into three layers by S31 on tri- Color Channels of R, B, G, extract respectively in target area it is each
Know the pixel value of pixel.
It is guarantee restoring area and continuity of the known region in structure in the present invention, it is necessary to ensure that what is calculated is preferential
Power is maximum in structural region numerical value, smaller in texture region numerical value, is also minimum in smooth region numerical value.In view of structural region
There is border, the characteristics of color distortion is larger, smooth region color does not almost have difference just can be distinguished very well using standard deviation
Structural region and smooth region.But texture region color there may be larger difference, so simple use standard deviation is carried out
Calculating can not distinguish texture region and structural region.Although and texture region color distortion is larger, being evenly distributed using it
The characteristics of it can also be made a distinction with structural region.
S32 calculates each layer pixel value standard deviation S in target area respectivelym(i, j), calculation formula is as follows:
Wherein, Pm(i, j) is m layers of target area image, and the i-th row, jth arranges point pixel value, and n is in target area image
The number of known pixels point, as the preferred embodiments of the invention, n takes 9, E (i, j) to be the pixel period of m layers of correspondence position
Hope, i.e.,:
Each layer pixel value standard deviation of S33 integration objective area images, calculating obtains each pixel of architectural difference matrix correspondence
The value S (i, j) of position, calculation formula is as follows:
Because smooth region standard deviation is possible to integrally Jia 1 in 0, therefore formula, obtained value is stored in structure poor
The corresponding position of different matrix.Structural region and the corresponding matrix numerical value of texture region are larger in image, and smooth region is close to 1.
Relatively, edge point value can be bigger than around putting close texture region numerical value, but not necessarily than not around it
The value of other texture regions is big.The value of texture region and borderline region can not be distinguished well, if so simple uses square
The corresponding value of battle array is obviously unreasonable as priority.Have clearly a need for further processing.As the present invention preferably embodiment, with
Less image block is unit around complex point to be repaired, builds priority condition structural coefficient.
S4 extracts a certain size the window centered on any point p to be repaired on the border of area to be repaired in target area
Mouth region domain, calculates complex point p to be repaired marginal texture coefficient T (p), and calculation formula is as follows:
Wherein, Ns(p) it is window area, | Ns(p) | for known element number, V in window areap,kIt is the neighbour of complex point to be repaired
Domain point normalization result, is defined as:
Wherein Z (p) is normaliztion constant, is defined as:
In order to ensure that structural region priority valve is maximum, from the point in the rectangular window centered on complex point to be repaired in a matrix
Corresponding all values being not zero, definition calculates the condition of priority.
In view of the difference of structural region and texture region is to contain border in structural region, the corresponding value in border compares surrounding values
It is big, in order to protrude this feature, and allow because texture difference causes the different region of standard deviation to be in close reparation successively
Position, the present invention is then squared using being normalized to putting the corresponding value in architectural difference matrix around point to be repaired
The mode differential expression of sum.It so can be very good specification configuration region and texture region.
The value of quadratic sum can with the larger value of element in quadratic sum number reduce and increase, and structural region compared to
The difference of texture region is that only fewer parts is the larger fringe region of value, can ground differentiation texture region and structural area with this
Domain.Although different texture regional value differs greatly, the value difference in identical texture region is different smaller, thus to target point around
Point is normalized, and can make preferential position of all texture regions in close preferential position less than structural region.
With the increase of summation element number, numerical value can diminish, and influence the stability of priority value.Therefore ginseng need to be multiplied by
Number with element of summing is balanced.So providing complex point p to be repaired structural coefficient T (p).
As the present invention preferably embodiment, in units of less image block around complex point to be repaired, priority is built
Structural coefficient.It is texture region or structural region that the value, which can distinguish band and repair block, but if structural region but can not
The center for ensuring multiblock to be repaired is the junction section of pattern.Because when image block is in the same area and wherein containing identical
During the junction section of length, obtained value can be close.It is small both at one around complex point to be repaired to such as Fig. 2 a and Fig. 2 b
Region, it is clear that Fig. 2 a belong to texture region, and Fig. 2 b have obvious architectural feature, and first repair pattern junction section in figure b,
More preferable repairing effect can be obtained.Therefore the difference of the one less known region in multiblock center to be repaired is added in priority
It is worth (because when repairing image, central point is located at breakage image edge) to amplify the point in proximate region closer to border center.
S5 calculates standard deviation criteria Ss of the complex point p to be repaired in the window area centered on itp(p) and confidence level ginseng
Number C (p), wherein:
Standard deviation criteria Sp(p) calculation formula is:
Confidence level parameter C (p) is the ratio of known pixels point number and whole pixel numbers in window area.
It is to be repaired that S6 calculates this according to marginal texture coefficient, complex point standard deviation criteria to be repaired and pixel confidence parameter to be repaired
The priority valve P (p) of point, calculation formula is P (p)=T (p) C (p) Sp(p);
S7 repeat step S5-S6, until all complex points to be repaired on traversal area to be repaired border, obtain corresponding priority
Set.
S8 repairs the window area centered on the maximum complex point to be repaired of priority valve, and update area to be repaired border and
Architectural difference matrix.
S9 repeat step S4-S8, until area to be repaired repairs complete.
Present invention also offers a kind of image repair method based on above-mentioned precedence algorithm, weight is built according to provincial characteristics
The architectural difference matrix computations priority used again, that is, ensure that the continuity of picture structure part, while also reducing image
Operand in repair process.As shown in figure 3, its step includes:
Step 1, reading restored image to be repaired, and determine area to be repaired boundary point;
Step 2, the calculating each corresponding priority valve of complex point to be repaired in area to be repaired border;
Step 3, the complex point to be repaired for searching priority valve maximum, and repair the target area figure centered on this complex point to be repaired
As block;
Step 4, refreshing architectural difference matrix, judge whether the region do not repaired also, if so, then performing step 2- steps
Rapid 3, until image repair is complete.
Progress is done to precedence algorithm of the present invention and image mending method with reference to specific embodiment further
Explanation.This algorithm is based on Windows7 systems, Visual.Studio2013+OpenCV v2.4 and is used as experiment porch.Embodiment
In the Criminisi algorithms of the prior art for not considering structural coefficient algorithm as a comparison is respectively adopted, with invent it is described calculate
Method is compared, and is represented as shown in figure 4, repairing (a) row in the comparing result of image, figure for two methods without the original of breakage
Image, (b) row represent the damaged image for needing to be repaired, and (c) row represent the effect repaired using contrast algorithm, (d)
Row represent the effect repaired using this algorithm.
As can be seen from Figure, at piece image forehead, there is open defect in contrast algorithm, and inventive algorithm
Preferably picture can be restored;There is more serious defect at bridge opening for the second width image comparison algorithmic method, be
Because the unstability of reparation order, caused without the junction section for first repairing bridge opening bottom and sidepiece, and this paper algorithm
Preferably overcome this problem.There is obvious defect on the wing of aircraft for the 3rd width picture contrast algorithmic method,
Because the order of priority is not reasonable, cause to reuse what single match block was caused, this algorithm is by using area
Domain characteristic preferably builds the sequencing of priority, therefore does not occur problems, the 4th width figure and above-mentioned situation class
Seemingly.
For the using effect of two kinds of algorithms of more objective embodiment, using Y-PSNR PSNR (Peak Singe To
Noise Ratio) (unit decibel) measure the difference of image and original image after repairing, and the value of Y-PSNR is bigger, table
Show that the image after repairing and original undamaged image difference are smaller, the effect of reparation is better.The application algorithm and contrast algorithm
Y-PSNR it is as shown in table 1:
This patent algorithm of table 1 is compared with contrast arithmetic result
By table 1, inventive algorithm performance relatively contrasts algorithm and increased, and this is due to special using the region of picture
Property come build priority determine repair sequencing, optimize matching criterior and add stability.And then good holding structure
Partial continuity and it is not likely to produce error hiding phenomenon.
The invention discloses a kind of precedence algorithm built based on image border provincial characteristics to be repaired.The algorithm is first
Architectural difference matrix is built according to the provincial characteristics on area to be repaired border, edge to be repaired is built further according to architectural difference matrix
Structural coefficient, finally according to the marginal texture coefficient priority resolution to be repaired of structure, determines fill order.Gained priority is steady
It is fixed reliable, and the calculating of the priority based on architectural difference matrix can save substantial amounts of operation time, replicability is strong.
Finally it should be noted that:Various embodiments above is merely illustrative of the technical solution of the present invention, rather than its limitations;To the greatest extent
The present invention is described in detail with reference to foregoing embodiments for pipe, it will be understood by those within the art that:Its according to
The technical scheme described in foregoing embodiments can so be modified, or which part or all technical characteristic are entered
Row equivalent substitution;And these modifications or replacement, the essence of appropriate technical solution is departed from various embodiments of the present invention technology
The scope of scheme.
Claims (6)
1. a kind of precedence algorithm based on architectural difference Yu marginal texture coefficient to be repaired, it is characterised in that its step includes:
S1 reads damaged image, and marks area to be repaired;
S2 is extracted include area to be repaired border including, and the certain target area of size, and extract in target area it is all
Know pixel information;
S3 is according to the known pixels point information architecture architectural difference matrix of extraction;
S4 calculates area to be repaired border complex point p to be repaired marginal texture coefficient T (p);
S5 calculates standard deviation criteria Ss of the complex point p to be repaired in the window area centered on itpAnd confidence level parameter C (p)
(p);
S6 calculates this complex point to be repaired according to marginal texture coefficient, complex point standard deviation criteria to be repaired and pixel confidence parameter to be repaired
Priority valve P (p), calculation formula is P (p)=T (p) C (p) Sp(p);
S7 repeat step S5-S6, until all complex points to be repaired on traversal area to be repaired border, obtain corresponding priority collection
Close;
S8 repairs the window area centered on the maximum complex point to be repaired of priority valve, and updates area to be repaired border and structure
Difference matrix;
S9 repeat step S4-S8, until area to be repaired repairs complete.
2. the precedence algorithm according to claim 1 based on architectural difference Yu marginal texture coefficient to be repaired, its feature
It is, the step of step S3 builds architectural difference matrix includes:
Breakage image is accordingly divided into three layers by S31 on tri- Color Channels of R, B, G, and each known picture in target area is extracted respectively
The pixel value of vegetarian refreshments;
S32 calculates each layer pixel value standard deviation S in target area respectivelym(i, j), calculation formula is as follows:
<mrow>
<msub>
<mi>S</mi>
<mi>m</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>,</mo>
<mi>j</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<msqrt>
<mrow>
<mfrac>
<mn>1</mn>
<mi>n</mi>
</mfrac>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
<mrow>
<mi>i</mi>
<mo>+</mo>
<mn>1</mn>
</mrow>
</munderover>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>j</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
<mrow>
<mi>j</mi>
<mo>+</mo>
<mn>1</mn>
</mrow>
</munderover>
<msup>
<mrow>
<mo>&lsqb;</mo>
<msub>
<mi>P</mi>
<mi>m</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>,</mo>
<mi>j</mi>
<mo>)</mo>
</mrow>
<mo>-</mo>
<mi>E</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>,</mo>
<mi>j</mi>
<mo>)</mo>
</mrow>
<mo>&rsqb;</mo>
</mrow>
<mn>2</mn>
</msup>
</mrow>
</msqrt>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
</mrow>
Wherein, Pm(i, j) is m layers of target area image, and the i-th row, jth arranges point pixel value, and n is known in target area image
The number of pixel, E (i, j) is that the pixel of m layers of correspondence position is expected, i.e.,:
<mrow>
<msub>
<mi>E</mi>
<mi>m</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>,</mo>
<mi>j</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<mi>n</mi>
</mfrac>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
<mrow>
<mi>i</mi>
<mo>+</mo>
<mn>1</mn>
</mrow>
</munderover>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>j</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
<mrow>
<mi>j</mi>
<mo>+</mo>
<mn>1</mn>
</mrow>
</munderover>
<msub>
<mi>P</mi>
<mi>m</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>,</mo>
<mi>j</mi>
<mo>)</mo>
</mrow>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>2</mn>
<mo>)</mo>
</mrow>
<mo>;</mo>
</mrow>
Each layer pixel value standard deviation of S33 integration objective area images, calculating obtains each pixel position of architectural difference matrix correspondence
Value S (i, j), calculation formula is as follows:
<mrow>
<mi>S</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>,</mo>
<mi>j</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<mn>3</mn>
</mfrac>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>m</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mn>3</mn>
</munderover>
<msub>
<mi>S</mi>
<mi>m</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>,</mo>
<mi>j</mi>
<mo>)</mo>
</mrow>
<mo>+</mo>
<mn>1</mn>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>3</mn>
<mo>)</mo>
</mrow>
<mo>.</mo>
</mrow>
3. the precedence algorithm according to claim 1 based on architectural difference Yu marginal texture coefficient to be repaired, its feature
It is, the calculation formula that step S4 calculates area to be repaired marginal texture coefficient T (p) is as follows:
<mrow>
<mi>T</mi>
<mrow>
<mo>(</mo>
<mi>p</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<msqrt>
<mrow>
<mo>&lsqb;</mo>
<msub>
<mi>&Sigma;</mi>
<mrow>
<mi>k</mi>
<mo>&Element;</mo>
<msub>
<mi>N</mi>
<mi>s</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>p</mi>
<mo>)</mo>
</mrow>
</mrow>
</msub>
<msubsup>
<mi>V</mi>
<mrow>
<mi>p</mi>
<mo>,</mo>
<mi>k</mi>
</mrow>
<mn>2</mn>
</msubsup>
<mo>&rsqb;</mo>
<mo>&CenterDot;</mo>
<mo>|</mo>
<msub>
<mi>N</mi>
<mi>s</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>p</mi>
<mo>)</mo>
</mrow>
<mo>|</mo>
</mrow>
</msqrt>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>4</mn>
<mo>)</mo>
</mrow>
</mrow>
Wherein, Ns(p) it is window area, | Ns(p) | for known element number, V in window areap,kIt is neighborhood of a point point to be repaired
Result is normalized, is defined as:
<mrow>
<msub>
<mi>V</mi>
<mrow>
<mi>p</mi>
<mo>,</mo>
<mi>k</mi>
</mrow>
</msub>
<mo>=</mo>
<mfrac>
<mrow>
<mi>S</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>,</mo>
<mi>j</mi>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<mi>Z</mi>
<mrow>
<mo>(</mo>
<mi>p</mi>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>5</mn>
<mo>)</mo>
</mrow>
</mrow>
Wherein Z (p) is normaliztion constant, is defined as:
<mrow>
<mi>Z</mi>
<mrow>
<mo>(</mo>
<mi>p</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<mrow>
<mo>|</mo>
<msub>
<mi>N</mi>
<mi>s</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>p</mi>
<mo>)</mo>
</mrow>
<mo>|</mo>
</mrow>
</mfrac>
<msub>
<mi>&Sigma;</mi>
<mrow>
<mi>k</mi>
<mo>&Element;</mo>
<msub>
<mi>N</mi>
<mi>s</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>p</mi>
<mo>)</mo>
</mrow>
</mrow>
</msub>
<mi>S</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>,</mo>
<mi>j</mi>
<mo>)</mo>
</mrow>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>6</mn>
<mo>)</mo>
</mrow>
<mo>.</mo>
</mrow>
4. the precedence algorithm according to claim 1 based on architectural difference Yu marginal texture coefficient to be repaired, its feature
It is, the poor parameter S of the step S5 Playsp(p) calculation formula is:
<mrow>
<msub>
<mi>S</mi>
<mi>p</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>p</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<mn>3</mn>
</mfrac>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>m</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mn>3</mn>
</munderover>
<msqrt>
<mrow>
<mfrac>
<mn>1</mn>
<mi>n</mi>
</mfrac>
<msubsup>
<mi>&Sigma;</mi>
<mrow>
<mi>i</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
<mrow>
<mi>i</mi>
<mo>+</mo>
<mn>1</mn>
</mrow>
</msubsup>
<msubsup>
<mi>&Sigma;</mi>
<mrow>
<mi>j</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
<mrow>
<mi>j</mi>
<mo>+</mo>
<mn>1</mn>
</mrow>
</msubsup>
<msup>
<mrow>
<mo>(</mo>
<msub>
<mi>p</mi>
<mi>m</mi>
</msub>
<mo>(</mo>
<mi>i</mi>
<mo>,</mo>
<mi>j</mi>
<mo>)</mo>
<mo>-</mo>
<msub>
<mi>E</mi>
<mi>m</mi>
</msub>
<mo>(</mo>
<mi>p</mi>
<mo>)</mo>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
</mrow>
</msqrt>
<mo>+</mo>
<mn>1</mn>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>7</mn>
<mo>)</mo>
</mrow>
<mo>.</mo>
</mrow>
5. the precedence algorithm according to claim 1 based on architectural difference Yu marginal texture coefficient to be repaired, its feature
It is, confidence level parameter C (p) is the ratio of known pixels point number and whole pixel numbers in window area in the step S5
Value.
6. a kind of image repair method based on precedence algorithm described in claim 1, it is characterised in that its step includes:
Step 1, reading restored image to be repaired, and determine area to be repaired boundary point;
Step 2, the calculating each corresponding priority valve of complex point to be repaired in area to be repaired border;
Step 3, the complex point to be repaired for searching priority valve maximum, and repair the target area image centered on this complex point to be repaired
Block;
Step 4, refreshing architectural difference matrix, judge whether the region do not repaired also, if so, step 2- steps 3 are then performed,
Until image repair is complete.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710438763.8A CN107194897B (en) | 2017-06-12 | 2017-06-12 | A kind of precedence algorithm and image repair method based on architectural difference Yu marginal texture coefficient to be repaired |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710438763.8A CN107194897B (en) | 2017-06-12 | 2017-06-12 | A kind of precedence algorithm and image repair method based on architectural difference Yu marginal texture coefficient to be repaired |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107194897A true CN107194897A (en) | 2017-09-22 |
CN107194897B CN107194897B (en) | 2019-07-23 |
Family
ID=59876451
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710438763.8A Expired - Fee Related CN107194897B (en) | 2017-06-12 | 2017-06-12 | A kind of precedence algorithm and image repair method based on architectural difference Yu marginal texture coefficient to be repaired |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107194897B (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111915507A (en) * | 2020-06-19 | 2020-11-10 | 西安理工大学 | Repair method for falling-off of mural |
CN112330643A (en) * | 2020-11-10 | 2021-02-05 | 国网湖北省电力有限公司宜昌供电公司 | Secondary equipment state identification method based on sparse representation image restoration |
CN116153476A (en) * | 2023-01-04 | 2023-05-23 | 深圳市德钰医疗器械有限公司 | Digital false tooth intelligent processing system and method based on medical internet of things |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20050028741A (en) * | 2003-09-19 | 2005-03-23 | 한국전자통신연구원 | Apparatus and method for yielding wavelets coefficient |
CN101504764A (en) * | 2009-01-19 | 2009-08-12 | 西安交通大学 | Ancient architecture oil-decorative painting virtual repairing system and method |
CN106023089A (en) * | 2016-01-19 | 2016-10-12 | 河南理工大学 | Image restoration method based on block matching |
CN106504203A (en) * | 2016-09-29 | 2017-03-15 | 河南理工大学 | A kind of Block- matching image repair method of application screening strategy |
CN106530247A (en) * | 2016-11-04 | 2017-03-22 | 天津大学 | Multi-scale image restoring algorithm based on structure information |
-
2017
- 2017-06-12 CN CN201710438763.8A patent/CN107194897B/en not_active Expired - Fee Related
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20050028741A (en) * | 2003-09-19 | 2005-03-23 | 한국전자통신연구원 | Apparatus and method for yielding wavelets coefficient |
CN101504764A (en) * | 2009-01-19 | 2009-08-12 | 西安交通大学 | Ancient architecture oil-decorative painting virtual repairing system and method |
CN106023089A (en) * | 2016-01-19 | 2016-10-12 | 河南理工大学 | Image restoration method based on block matching |
CN106504203A (en) * | 2016-09-29 | 2017-03-15 | 河南理工大学 | A kind of Block- matching image repair method of application screening strategy |
CN106530247A (en) * | 2016-11-04 | 2017-03-22 | 天津大学 | Multi-scale image restoring algorithm based on structure information |
Non-Patent Citations (1)
Title |
---|
王新年 等: "基于几何距离的Criminisi图像修复算法", 《计算机工程与设计》 * |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111915507A (en) * | 2020-06-19 | 2020-11-10 | 西安理工大学 | Repair method for falling-off of mural |
CN111915507B (en) * | 2020-06-19 | 2023-10-17 | 西安理工大学 | Repairing method for wall painting falling off |
CN112330643A (en) * | 2020-11-10 | 2021-02-05 | 国网湖北省电力有限公司宜昌供电公司 | Secondary equipment state identification method based on sparse representation image restoration |
CN112330643B (en) * | 2020-11-10 | 2023-02-07 | 国网湖北省电力有限公司宜昌供电公司 | Secondary equipment state identification method based on sparse representation image restoration |
CN116153476A (en) * | 2023-01-04 | 2023-05-23 | 深圳市德钰医疗器械有限公司 | Digital false tooth intelligent processing system and method based on medical internet of things |
CN116153476B (en) * | 2023-01-04 | 2024-01-26 | 英德市优越义齿技术有限公司 | Digital false tooth intelligent processing system and method based on medical internet of things |
Also Published As
Publication number | Publication date |
---|---|
CN107194897B (en) | 2019-07-23 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Ni et al. | Out-of-core bundle adjustment for large-scale 3d reconstruction | |
CN105513064B (en) | A kind of solid matching method based on image segmentation and adaptive weighting | |
CN100458846C (en) | A method of image restoration | |
WO2021203795A1 (en) | Pancreas ct automatic segmentation method based on saliency dense connection expansion convolutional network | |
CN107451997A (en) | A kind of automatic identifying method of the welding line ultrasonic TOFD D scanning defect types based on deep learning | |
CN106683173A (en) | Method of improving density of three-dimensional reconstructed point cloud based on neighborhood block matching | |
CN107194897A (en) | A kind of precedence algorithm and image repair method based on architectural difference Yu marginal texture coefficient to be repaired | |
CN103839223A (en) | Image processing method and image processing device | |
CN109493320A (en) | Method for extracting remote sensing image road and system, storage medium, electronic equipment based on deep learning | |
CN108776772A (en) | Across the time building variation detection modeling method of one kind and detection device, method and storage medium | |
Yao | Damaged region filling by improved criminisi image inpainting algorithm for thangka | |
CN109859114A (en) | Three-dimensional point cloud restorative procedure based on local flatness and non-local similitude | |
CN108596975A (en) | A kind of Stereo Matching Algorithm for weak texture region | |
Poon et al. | Efficient interactive 3D Livewire segmentation of complex objects with arbitrary topology | |
CN104680492B (en) | Image repair method based on composition of sample uniformity | |
CN103065351A (en) | Binocular three-dimensional reconstruction method | |
Cao et al. | Restoration of an ancient temple mural by a local search algorithm of an adaptive sample block | |
CN103544695B (en) | A kind of efficiently based on the medical image cutting method of game framework | |
CN110544253B (en) | Fabric flaw detection method based on image pyramid and column template | |
CN108629809B (en) | Accurate and efficient stereo matching method | |
CN108052909A (en) | A kind of thin fibrous cap patch automatic testing method and device based on cardiovascular OCT images | |
Schmidt et al. | Forest point processes for the automatic extraction of networks in raster data | |
CN115410189A (en) | Complex scene license plate detection method | |
CN107833191A (en) | Improvement Criminisi algorithms based on image local information | |
CN107369138B (en) | Image optimization display method based on high-order statistical model |
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 | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20190723 Termination date: 20200612 |
|
CF01 | Termination of patent right due to non-payment of annual fee |