CN108171675A - A kind of image repair method and device based on separation Bregman iteration optimizations - Google Patents
A kind of image repair method and device based on separation Bregman iteration optimizations Download PDFInfo
- Publication number
- CN108171675A CN108171675A CN201810218680.2A CN201810218680A CN108171675A CN 108171675 A CN108171675 A CN 108171675A CN 201810218680 A CN201810218680 A CN 201810218680A CN 108171675 A CN108171675 A CN 108171675A
- Authority
- CN
- China
- Prior art keywords
- image
- iteration
- object function
- separation
- formula
- 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
- 238000005457 optimization Methods 0.000 title claims abstract description 44
- 238000000034 method Methods 0.000 title claims abstract description 41
- 238000000926 separation method Methods 0.000 title claims abstract description 41
- 230000008439 repair process Effects 0.000 title claims abstract description 30
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 27
- 238000004364 calculation method Methods 0.000 claims abstract description 17
- 239000000284 extract Substances 0.000 claims description 14
- 230000008569 process Effects 0.000 claims description 10
- 238000001914 filtration Methods 0.000 claims description 8
- 238000013480 data collection Methods 0.000 claims description 7
- 238000009795 derivation Methods 0.000 claims description 6
- 230000000694 effects Effects 0.000 abstract description 9
- 230000006870 function Effects 0.000 description 66
- 238000005516 engineering process Methods 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 230000008878 coupling Effects 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 238000009499 grossing Methods 0.000 description 2
- 238000011084 recovery Methods 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- 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
- G06T5/10—Image enhancement or restoration using non-spatial domain filtering
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
Abstract
A kind of image repair method and device based on separation Bregman iteration optimizations provided by the invention, wherein method include：The object function based on the desired image restoration model of block likelihood logarithm is established, by introducing intermediate auxiliary variableObject function is converted into finite form；Call separation Bregman iteration optimization algorithms that the object function of finite form is divided into about ziWith the iterative calculation formula of X, the z obtained when reaching maximum iteration is calculatediAnd X；X when being up to maximum iteration is as the image output after repairing.The present invention is iterated to calculate by introducing intermediate auxiliary variable and dividing object function with separation Bregman iteration optimization algorithms, can improve the convergence rate of image repair, while effectively improve the repairing effect of image.
Description
Technical field
The present invention relates to technical field of image processing more particularly to a kind of images based on separation Bregman iteration optimizations
Restorative procedure and device.
Background technology
Image repair is all a research hotspot of digital image processing field all the time.Image is during transmission
Can because it is various the problem of occur damaged, there is the phenomenon that degrades so as to cause image.Image repair is exactly to arrive according to the observation
Image information damaged area is repaired, the image after reconstruction is made to reach the visual effect close to artwork.
The distribution of its learning image prior is widely applied in the field of image restoration, such as image denoising, image
It repairs, image deblurring etc..In image restoration field, prior image model can be divided into following three kinds of models：Image local
Smoothing model, the non local self similarity model of image, image sparse represent model.Image local smoothing model is utilized in local neck
Existing correlation and continuity model between image pixel in domain.Wherein typical model has Total Variation, autoregression mould
Type etc..The non local self similarity model of image finds the image block for having similar structure to degraded image block in the picture, utilizes this
The pixel of image block restores degraded image block, so as to restoring entire image.Wherein typical model has non local averaging model,
Non local Total Variation etc..Image sparse represents that model is described picture signal and can be approached using a small amount of basic function
It represents.These set of basis function collectively form a dictionary.
It is worth noting that, above-mentioned model is all to use to be modeled based on image block processing thought, degraded image point
A series of image block is cut into, degraded image block is restored using block prior distribution.Experiment is it has been shown that based on block priori
Image restoration is than directly much simpler to the recovery of entire image priori, and image restoration effect is also relatively good.Danie et al.
It proposes study and damaged image is restored using the priori of image block, while calculated with half secondary split method
Method optimizes, different prioris, the reparation of damaged image is generated different as a result, the experiment of Danie shows to select GMM first
It is more preferable to the repairing effect of damaged image than ICA priori to test knowledge.
The mathematical model of block log-likelihood function can be expressed as by formula：Wherein PiRepresent from
The operation of i-th piece of image block is extracted in image x.It is worth noting that PiWhat x was extracted is a series of figures to lie overlapping one another
As block.logp(PiX) logarithm of i-th piece of image block possibility predication under given prior distribution p is represented.
The basic thought of EPLL is the likelihood logarithm for maximizing image block.Assuming that by a sub-picture be divided into one group it is mutual
The image block of overlapping, estimates the likelihood logarithm of every piece of image block according to EPLL, and Danie Zoran think every piece of image block
Likelihood logarithm is bigger, is more conducive to restore breakage image, and recovery effect is better.
The image restoration model for it is expected (EPLL) based on block likelihood logarithm is represented by：
Wherein Y represents the degraded image information that actual observation arrives, and X represents to rebuild obtained image, and A represents that linearity moves back
Change function, | | AX-Y | |2Represent image fidelity item.
Block likelihood logarithm based on the division of half quadratic power it is expected the image restoration model of (EPLL) by introducing image block number
According to collectionAs intermediate auxiliary variable, object function can be converted to following form：
Block likelihood logarithm based on half quadratic power splitting algorithm it is expected (EPLL) image restoration model, has been applied to figure
As process field, and obtain very well results.The drawback is that power consumption is big, convergence rate is slow, and image repair effect is not fine.
Invention content
The present invention provides a kind of image repair method and device based on separation Bregman iteration optimizations, for solving
The technical issues of traditional image repair technology power consumption is big, and convergence rate is slow, and image repair effect is not fine.
A kind of image repair method based on separation Bregman iteration optimizations provided by the invention, including：
S1：The object function based on the desired image restoration model of block likelihood logarithm is established, is become by introducing intermediate auxiliary
AmountObject function is converted into finite form；
S2：Call separation Bregman iteration optimization algorithms that the object function of finite form is divided into about ziWith changing for X
For calculation formula, the z obtained when reaching maximum iteration is calculatediAnd X；
S3：X when being up to maximum iteration is as the image output after repairing；
Wherein,For image block data collection, zi=PiThe piecemeal that X, wherein P are image X extracts function, PiFor image X's
I-th piece of image block extracts function, and the initial value of image X is image to be repaired, and image X when reaching maximum iteration is repaiies
Image after multiple.
Preferably, the step S1 includes：
Establishing the object function based on the desired image restoration model of block likelihood logarithm is：
Wherein, wherein Y represents the degraded image information that actual observation arrives, and X represents to rebuild obtained image, and A represents linearity
Degenrate function, | | AX-Y | |2Represent image fidelity item；
By introducing intermediate auxiliary variableObject function is converted into finite form is：
The object function of finite form is become by Lagrange multiplier：
Wherein, β is preset penalty term parameter.
Preferably, the step S2 includes：
Separation Bregman iteration optimization algorithms are called, introduce iterative parameter variable b, the object function that step S1 is exported
It is decomposed into the first iterative formula, secondary iteration formula and third iterative formula；
First iterative formula is：
The secondary iteration formula is：
The third iterative formula is：
Maximum iteration is set, and repetition calculates z successivelyi, X and b, until reaching maximum iteration；
Wherein, β is preset penalty term parameter, and Y is degraded image.
Preferably, z in the step S2iCalculating process be：
Enable ri=Pi(X(k)-b(k)), it substitutes into the first iterative formula and obtains：
Utilize given image data setNew image is obtained by Wiener filtering algorithm with Gaussian prior model
Data set
Preferably, the calculating process of X is in the step S2：
To the derivation of secondary iteration formula and its derivative is made to be obtained for 0：
By what is solvedThe formula, which calculates, obtains image X；
Wherein, Y is degraded image.
Preferably, the solution procedure of b is in the step S2：
By arriving for solutionThird iterative formula is substituted into X, and b is calculated.
A kind of image fixing apparatus based on separation Bregman iteration optimizations provided by the invention, including：
Object function module for establishing the object function based on the desired image restoration model of block likelihood logarithm, passes through
Introduce intermediate auxiliary variableObject function is converted into finite form；
Module is iterated to calculate, for separation Bregman iteration optimization algorithms to be called to divide the object function of finite form
Into about ziWith the iterative calculation formula of X, the z obtained when reaching maximum iteration is calculatediAnd X；
Output module, the X for being up to during maximum iteration are exported as the image after repairing；
Wherein,For image block data collection, zi=PiThe piecemeal that X, wherein P are image X extracts function, PiFor image X's
I-th piece of image block extracts function, and the initial value of image X is image to be repaired, and image X when reaching maximum iteration is repaiies
Image after multiple.
Preferably, the object function module includes：
Object function establishes unit, for establishing the object function based on the desired image restoration model of block likelihood logarithm
For：Wherein, wherein Y represents the degraded image information that actual observation arrives, and X is represented
Obtained image is rebuild, A represents linearity degenrate function, | | AX-Y | |2Represent image fidelity item；
Auxiliary variable unit introduces intermediate auxiliary variable for passing throughObject function is converted into finite form is：
Lagrangian unit becomes the object function of finite form for passing through Lagrange multiplier：
Wherein, β is preset penalty term parameter.
Preferably, the iterative calculation module includes：
Object function resolving cell detaches Bregman iteration optimization algorithms for calling, introduces iterative parameter variable b, will
The object function of object function module output is decomposed into the first iterative formula, secondary iteration formula and third iterative formula；
First iterative formula is：
The secondary iteration formula is：
The third iterative formula is：
Unit is iterated to calculate, for setting maximum iteration, repetition calculates z successivelyi, X and b, change until reaching maximum
Generation number；
Wherein, β is preset penalty term parameter, and Y is degraded image.
Preferably, the iterative calculation unit includes ziComputation subunit, ziComputation subunit is used for：
Enable ri=Pi(X(k)-b(k)), it substitutes into the first iterative formula and obtains：
Utilize given image data setNew image is obtained by Wiener filtering algorithm with Gaussian prior model
Data set
Preferably, the iterative calculation unit further includes X computation subunits, and X computation subunits are used for：
To the derivation of secondary iteration formula and its derivative is made to be obtained for 0：
By what is solvedThe formula, which calculates, obtains image X；
Wherein, Y is degraded image.
Preferably, the iterative calculation unit further includes b computation subunits, and b computation subunits are used for：
By arriving for solutionThird iterative formula is substituted into X, and b is calculated.
As can be seen from the above technical solutions, the present invention has the following advantages：
A kind of image repair method and device based on separation Bregman iteration optimizations provided by the invention, wherein method
Including：S1：The object function based on the desired image restoration model of block likelihood logarithm is established, by introducing intermediate auxiliary variableObject function is converted into finite form；S2：Separation Bregman iteration optimization algorithms are called by the target letter of finite form
Number is divided into about ziWith the iterative calculation formula of X, the z obtained when reaching maximum iteration is calculatediAnd X；S3：It is up to most
X during big iterations is as the image output after repairing.The present invention is by introducing intermediate auxiliary variable and with separation
Bregman iteration optimization algorithms, which divide object function, to be iterated to calculate, and can improve the convergence rate of image repair, while effectively
Improve the repairing effect of image.
Description of the drawings
In order to illustrate more clearly about the embodiment of the present invention or technical scheme of the prior art, to embodiment or will show below
There is attached drawing needed in technology description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention, for those of ordinary skill in the art, without having to pay creative labor, may be used also
To obtain other attached drawings according to these attached drawings.
Fig. 1 is an a kind of implementation of image repair method based on separation Bregman iteration optimizations provided by the invention
The schematic diagram of example.
Specific embodiment
The present invention provides a kind of image repair method and device based on separation Bregman iteration optimizations, for solving
The technical issues of traditional image repair technology power consumption is big, and convergence rate is slow, and image repair effect is not fine.
In order to make the invention's purpose, features and advantages of the invention more obvious and easy to understand, below in conjunction with the present invention
Attached drawing in embodiment is clearly and completely described the technical solution in the embodiment of the present invention, it is clear that disclosed below
Embodiment be only part of the embodiment of the present invention, and not all embodiment.Based on the embodiments of the present invention, this field
All other embodiment that those of ordinary skill is obtained without making creative work, belongs to protection of the present invention
Range.
Referring to Fig. 1, one of a kind of image repair method based on separation Bregman iteration optimizations provided by the invention
Embodiment, including：
101：The object function based on the desired image restoration model of block likelihood logarithm is established, is become by introducing intermediate auxiliary
AmountObject function is converted into finite form；
102：Call separation Bregman iteration optimization algorithms that the object function of finite form is divided into about ziWith X's
Formula is iterated to calculate, calculates the z obtained when reaching maximum iterationiAnd X；
103：X when being up to maximum iteration is as the image output after repairing；
Wherein,For image block data collection, zi=PiThe piecemeal that X, wherein P are image X extracts function, PiFor image X's
I-th piece of image block extracts function, and the initial value of image X is image to be repaired, and image X when reaching maximum iteration is repaiies
Image after multiple.
The present invention is by introducing intermediate auxiliary variable and dividing object function with separation Bregman iteration optimization algorithms
Iterative calculation, can improve the convergence rate of image repair, while effectively improve the repairing effect of image.
It should be noted that data setInclude several zi=PiX, i=1,2,3 ... N.
Further, step 101 includes：
Establishing the object function based on the desired image restoration model of block likelihood logarithm is：
Wherein, wherein Y represents the degraded image information that actual observation arrives, and X represents to rebuild obtained image, and A represents that linearity is degenerated
Function, | | AX-Y | |2Represent image fidelity item；
By introducing intermediate auxiliary variableObject function is converted into finite form is：
The object function of finite form is become by Lagrange multiplier：
Wherein, β is preset penalty term parameter.
Further, step 102 includes：
Separation Bregman iteration optimization algorithms are called, introduce iterative parameter variable b, the object function that step 101 is exported
It is decomposed into the first iterative formula, secondary iteration formula and third iterative formula；
First iterative formula is：
Secondary iteration formula is：
Third iterative formula is：
Maximum iteration is set, and repetition calculates z successivelyi, X and b, until reaching maximum iteration；
Wherein, β is preset penalty term parameter, and Y is degraded image.
Further, z in step 102iCalculating process be：
Enable ri=Pi(X(k)-b(k)), it substitutes into the first iterative formula and obtains：
Utilize given image data setNew image is obtained by Wiener filtering algorithm with Gaussian prior model
Data set
Further, the calculating process of X is in step 102：
To the derivation of secondary iteration formula and its derivative is made to be obtained for 0：
By what is solvedThe formula, which calculates, obtains image X；
Wherein, Y is degraded image.
Further, the solution procedure of b is in step 102：
By arriving for solutionThird iterative formula is substituted into X, and b is calculated.
It is to an a kind of reality of the image repair method based on separation Bregman iteration optimizations provided by the invention above
Example is applied to be described in detail, it below will be to a kind of image repair based on separation Bregman iteration optimizations provided by the invention
Another embodiment of method is described in detail.
A kind of another embodiment of the image repair method based on separation Bregman iteration optimizations, including：
1st, first by introducing an intermediate auxiliary variable so that object function is equivalently changed into following limited shape
Formula：
Specifically we enable zi=PiX, wherein, the piecemeal that P is image X extracts function, PiI-th piece of image block for image X
Extract function.We become object function with Lagrange multiplier in this way：
Bregman can be used to calculate it should be noted that Argmin is a kind of representation object function minimized
Method.
2nd, separation Bregman iteration optimization algorithms are called, introduce an iterative parameter variable b by above formula objective function Equation
It is divided into following three problems of iterative solution：
By setting iterations, z is obtained successivelyi, X, b value, until reaching maximum iteration.
Specifically, r is enabledi=Pi(X(k)-b(k)), then formula (1) becomes：
Utilize given image data setNew image is obtained by Wiener filtering algorithm with Gaussian prior model
Data setDetailed process is as follows：
In GMM model, the log-likelihood estimation for calculating given image block is represented by：
Wherein πkRepresent hybrid weight value in mixing portion, μkAnd ∑kMean value and covariance matrix are represented respectively.
First, the hybrid weight π of the pixel mixing portion of given noise image block is calculatedk=P (k | y), next, choosing
Select hybrid weight value k maximum in mixing portionmax=maxkπk, the Gaussian kernel classification of each image block is obtained, then with dimension
Filtering algorithm of receiving calculating is obtained by MAP estimation：
In formula
3rd, next, utilizing the image block data collection acquiredTo be updated to X.It is brought into solve in (2) and obtain：
Specifically, by the way that its derivative is made to be 0 to formula (2) derivation, X is obtained(k+1)Expression formula, then willSubstitute into expression
Formula, it is possible to obtain X(k+1)。
4th, the X and z solved, which is brought into solve in (3), obtains new b.
5th, by setting certain iterations, above-mentioned iterative step is repeated, until reaching maximum iteration, output is repaiied
Image X after multiple.
It below will be to an a kind of reality of the image fixing apparatus based on separation Bregman iteration optimizations provided by the invention
Example is applied to be described in detail.
A kind of one embodiment of image fixing apparatus based on separation Bregman iteration optimizations provided by the invention, packet
It includes：
Object function module for establishing the object function based on the desired image restoration model of block likelihood logarithm, passes through
Introduce intermediate auxiliary variableObject function is converted into finite form；
Module is iterated to calculate, for separation Bregman iteration optimization algorithms to be called to divide the object function of finite form
Into about ziWith the iterative calculation formula of X, the z obtained when reaching maximum iteration is calculatediAnd X；
Output module, the X for being up to during maximum iteration are exported as the image after repairing；
Wherein,For image block data collection, zi=PiThe piecemeal that X, wherein P are image X extracts function, PiFor image X's
I-th piece of image block extracts function, and the initial value of image X is image to be repaired, and image X when reaching maximum iteration is repaiies
Image after multiple.
Further, object function module includes：
Object function establishes unit, for establishing the object function based on the desired image restoration model of block likelihood logarithm
For：Wherein, wherein Y represents the degraded image information that actual observation arrives, and X is represented
Obtained image is rebuild, A represents linearity degenrate function, | | AX-Y | |2Represent image fidelity item；
Auxiliary variable unit introduces intermediate auxiliary variable for passing throughObject function is converted into finite form is：
Lagrangian unit becomes the object function of finite form for passing through Lagrange multiplier：
Wherein, β is preset penalty term parameter.
Further, iterative calculation module includes：
Object function resolving cell detaches Bregman iteration optimization algorithms for calling, introduces iterative parameter variable b, will
The object function of object function module output is decomposed into the first iterative formula, secondary iteration formula and third iterative formula；
First iterative formula is：
Secondary iteration formula is：
Third iterative formula is：
Unit is iterated to calculate, for setting maximum iteration, repetition calculates z successivelyi, X and b, change until reaching maximum
Generation number；
Wherein, β is preset penalty term parameter, and Y is degraded image.
Further, iterative calculation unit includes ziComputation subunit, ziComputation subunit is used for：
Enable ri=Pi(X(k)-b(k)), it substitutes into the first iterative formula and obtains：
Utilize given image data setNew image is obtained by Wiener filtering algorithm with Gaussian prior model
Data set
Further, iterative calculation unit further includes X computation subunits, and X computation subunits are used for：
To the derivation of secondary iteration formula and its derivative is made to be obtained for 0：
By what is solvedThe formula, which calculates, obtains image X；
Wherein, Y is degraded image.
Further, iterative calculation unit further includes b computation subunits, and b computation subunits are used for：
By arriving for solutionThird iterative formula is substituted into X, and b is calculated.
It is apparent to those skilled in the art that for convenience and simplicity of description, the system of foregoing description,
The specific work process of device and unit can refer to the corresponding process in preceding method embodiment, and details are not described herein.
In several embodiments provided herein, it should be understood that disclosed system, device and method can be with
It realizes by another way.For example, the apparatus embodiments described above are merely exemplary, for example, the unit
It divides, only a kind of division of logic function can have other dividing mode, such as multiple units or component in actual implementation
It may be combined or can be integrated into another system or some features can be ignored or does not perform.Another point, it is shown or
The mutual coupling, direct-coupling or communication connection discussed can be the indirect coupling by some interfaces, device or unit
It closes or communicates to connect, can be electrical, machinery or other forms.
The unit illustrated as separating component may or may not be physically separate, be shown as unit
The component shown may or may not be physical unit, you can be located at a place or can also be distributed to multiple
In network element.Some or all of unit therein can be selected according to the actual needs to realize the mesh of this embodiment scheme
's.
In addition, each functional unit in each embodiment of the present invention can be integrated in a processing unit, it can also
That each unit is individually physically present, can also two or more units integrate in a unit.Above-mentioned integrated list
The form that hardware had both may be used in member is realized, can also be realized in the form of SFU software functional unit.
If the integrated unit is realized in the form of SFU software functional unit and is independent product sale or uses
When, it can be stored in a computer read/write memory medium.Based on such understanding, technical scheme of the present invention is substantially
The part to contribute in other words to the prior art or all or part of the technical solution can be in the form of software products
It embodies, which is stored in a storage medium, is used including some instructions so that a computer
Equipment (can be personal computer, server or the network equipment etc.) performs the complete of each embodiment the method for the present invention
Portion or part steps.And aforementioned storage medium includes：USB flash disk, mobile hard disk, read-only memory (ROM, Read-Only
Memory), random access memory (RAM, Random Access Memory), magnetic disc or CD etc. are various can store journey
The medium of sequence code.
The above, the above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations；Although with reference to before
Embodiment is stated the present invention is described in detail, it will be understood by those of ordinary skill in the art that：It still can be to preceding
The technical solution recorded in each embodiment is stated to modify or carry out equivalent replacement to which part technical characteristic；And these
Modification is replaced, the spirit and scope for various embodiments of the present invention technical solution that it does not separate the essence of the corresponding technical solution.
Claims (10)
1. a kind of image repair method based on separation Bregman iteration optimizations, which is characterized in that including：
S1：The object function based on the desired image restoration model of block likelihood logarithm is established, by introducing intermediate auxiliary variableObject function is converted into finite form；
S2：Call separation Bregman iteration optimization algorithms that the object function of finite form is divided into about ziWith the iteration meter of X
Formula is calculated, calculates the z obtained when reaching maximum iterationiAnd X；
S3：X when being up to maximum iteration is as the image output after repairing；
Wherein,For image block data collection, zi=PiThe piecemeal that X, wherein P are image X extracts function, PiI-th for image X
Block image block extracts function, and the initial value of image X is image to be repaired, and image X when reaching maximum iteration is reparation
Image afterwards.
2. a kind of image repair method based on separation Bregman iteration optimizations according to claim 1, feature exist
In the step S1 includes：
Establishing the object function based on the desired image restoration model of block likelihood logarithm is：
Wherein, wherein Y represents the degraded image information that actual observation arrives, and X represents to rebuild obtained image, and A represents linearity
Degenrate function, | | AX-Y | |2Represent image fidelity item；
By introducing intermediate auxiliary variableObject function is converted into finite form is：
The object function of finite form is become by Lagrange multiplier λ：
Wherein, β is preset penalty term parameter.
3. a kind of image repair method based on separation Bregman iteration optimizations according to claim 1, feature exist
In the step S2 includes：
Separation Bregman iteration optimization algorithms are called, introduce iterative parameter variable b, the object function that step S1 is exported decomposes
For the first iterative formula, secondary iteration formula and third iterative formula；
First iterative formula is：
The secondary iteration formula is：
The third iterative formula is：
Maximum iteration is set, and repetition calculates z successivelyi, X and b, until reaching maximum iteration；
Wherein, β is preset penalty term parameter, and Y is degraded image.
4. a kind of image repair method based on separation Bregman iteration optimizations according to claim 3, feature exist
In z in the step S2iCalculating process be：
Enable ri=Pi(X(k)-b(k)), it substitutes into the first iterative formula and obtains：
Utilize preset image data setNew image data is obtained by Wiener filtering algorithm with Gaussian prior model
Collection
5. a kind of image repair method based on separation Bregman iteration optimizations according to claim 4, feature exist
In the calculating process of X is in the step S2：
To the derivation of secondary iteration formula and its derivative is made to be obtained for 0：
By what is solvedThe formula, which calculates, obtains image X；
Wherein, Y is degraded image.
6. a kind of image repair method based on separation Bregman iteration optimizations according to claim 4, feature exist
In the solution procedure of b is in the step S2：
By arriving for solutionThird iterative formula is substituted into X, and b is calculated.
7. a kind of image fixing apparatus based on separation Bregman iteration optimizations, which is characterized in that including：
Object function module for establishing the object function based on the desired image restoration model of block likelihood logarithm, passes through introducing
Intermediate auxiliary variableObject function is converted into finite form；
Module is iterated to calculate, for calling separation Bregman iteration optimization algorithms that the object function of finite form is divided into pass
In ziWith the iterative calculation formula of X, the z obtained when reaching maximum iteration is calculatediAnd X；
Output module, the X for being up to during maximum iteration are exported as the image after repairing；
Wherein,For image block data collection, zi=PiThe piecemeal that X, wherein P are image X extracts function, PiI-th for image X
Block image block extracts function, and the initial value of image X is image to be repaired, and image X when reaching maximum iteration is reparation
Image afterwards.
8. a kind of image fixing apparatus based on separation Bregman iteration optimizations according to claim 7, feature exist
In the object function module includes：
Object function establishes unit, is for establishing the object function based on the desired image restoration model of block likelihood logarithm：Wherein, wherein Y represents the degraded image information that actual observation arrives, and X represents to rebuild
Obtained image, A represent linearity degenrate function, | | AX-Y | |2Represent image fidelity item；
Auxiliary variable unit introduces intermediate auxiliary variable for passing throughObject function is converted into finite form is：
Lagrangian unit becomes the object function of finite form for passing through Lagrange multiplier：
Wherein, β is preset penalty term parameter.
9. a kind of image fixing apparatus based on separation Bregman iteration optimizations according to claim 7, feature exist
In the iterative calculation module includes：
Object function resolving cell detaches Bregman iteration optimization algorithms for calling, iterative parameter variable b is introduced, by target
The object function of function module output is decomposed into the first iterative formula, secondary iteration formula and third iterative formula；
First iterative formula is：
The secondary iteration formula is：
The third iterative formula is：
Unit is iterated to calculate, for setting maximum iteration, repetition calculates z successivelyi, X and b, until reaching greatest iteration time
Number；
Wherein, β is preset penalty term parameter, and Y is degraded image.
10. a kind of image repair method based on separation Bregman iteration optimizations according to claim 9, feature exist
In the iterative calculation unit includes ziComputation subunit is used for：
Enable ri=Pi(X(k)-b(k)), it substitutes into the first iterative formula and obtains：
Utilize given image data setNew image data is obtained by Wiener filtering algorithm with Gaussian prior model
Collection
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810218680.2A CN108171675B (en) | 2018-03-16 | 2018-03-16 | Image restoration method and device based on separation Bregman iterative optimization |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810218680.2A CN108171675B (en) | 2018-03-16 | 2018-03-16 | Image restoration method and device based on separation Bregman iterative optimization |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108171675A true CN108171675A (en) | 2018-06-15 |
| CN108171675B CN108171675B (en) | 2021-10-01 |
Family
ID=62511232
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810218680.2A Active CN108171675B (en) | 2018-03-16 | 2018-03-16 | Image restoration method and device based on separation Bregman iterative optimization |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108171675B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109658463A (en) * | 2018-11-28 | 2019-04-19 | 中国科学院自动化研究所 | Exciting tomography fluorescence imaging method based on non local full variation regularization |
| CN110298798A (en) * | 2019-06-20 | 2019-10-01 | 浙江工业大学 | A kind of image repair method based on the completion of low-rank tensor Yu discrete full variation |
| CN112561842A (en) * | 2020-12-07 | 2021-03-26 | 昆明理工大学 | Multi-source damaged image fusion and recovery combined implementation method based on dictionary learning |
| CN113139920A (en) * | 2021-05-12 | 2021-07-20 | 闽南师范大学 | Ancient book image restoration method, terminal device and storage medium |
| CN114022387A (en) * | 2021-11-11 | 2022-02-08 | 上海联影医疗科技股份有限公司 | Image optimization method and device based on total variation minimization and computer equipment |
| CN114331853B (en) * | 2020-09-30 | 2023-05-12 | 四川大学 | Single image restoration iteration framework based on target vector updating module |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104134196A (en) * | 2014-08-08 | 2014-11-05 | 重庆大学 | Split Bregman weight iteration image blind restoration method based on non-convex higher-order total variation model |
| CN105046664A (en) * | 2015-07-13 | 2015-11-11 | 广东工业大学 | Image denoising method based on self-adaptive EPLL algorithm |
-
2018
- 2018-03-16 CN CN201810218680.2A patent/CN108171675B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104134196A (en) * | 2014-08-08 | 2014-11-05 | 重庆大学 | Split Bregman weight iteration image blind restoration method based on non-convex higher-order total variation model |
| CN105046664A (en) * | 2015-07-13 | 2015-11-11 | 广东工业大学 | Image denoising method based on self-adaptive EPLL algorithm |
Non-Patent Citations (2)
| Title |
|---|
| BO DONG ET AL.: "Improved patch-based learning for image deblurring", 《MOBILE MULTIMEDIA/IMAGE PROCESSING, SECURITY, AND APPLICATIONS 2015》 * |
| 熊耀先 等: "基于图像块统计特性的EPLL遥感图像去噪方法", 《数字技术与应用》 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109658463A (en) * | 2018-11-28 | 2019-04-19 | 中国科学院自动化研究所 | Exciting tomography fluorescence imaging method based on non local full variation regularization |
| CN109658463B (en) * | 2018-11-28 | 2021-07-09 | 中国科学院自动化研究所 | Excitation fluorescence tomography method based on non-local total variation regularization |
| CN110298798A (en) * | 2019-06-20 | 2019-10-01 | 浙江工业大学 | A kind of image repair method based on the completion of low-rank tensor Yu discrete full variation |
| CN114331853B (en) * | 2020-09-30 | 2023-05-12 | 四川大学 | Single image restoration iteration framework based on target vector updating module |
| CN112561842A (en) * | 2020-12-07 | 2021-03-26 | 昆明理工大学 | Multi-source damaged image fusion and recovery combined implementation method based on dictionary learning |
| CN112561842B (en) * | 2020-12-07 | 2022-12-09 | 昆明理工大学 | Multi-source damaged image fusion and recovery combined implementation method based on dictionary learning |
| CN113139920A (en) * | 2021-05-12 | 2021-07-20 | 闽南师范大学 | Ancient book image restoration method, terminal device and storage medium |
| CN113139920B (en) * | 2021-05-12 | 2023-05-12 | 闽南师范大学 | Ancient book image restoration method, terminal equipment and storage medium |
| CN114022387A (en) * | 2021-11-11 | 2022-02-08 | 上海联影医疗科技股份有限公司 | Image optimization method and device based on total variation minimization and computer equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108171675B (en) | 2021-10-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108171675A (en) | A kind of image repair method and device based on separation Bregman iteration optimizations | |
| CN109410261B (en) | Monocular image depth estimation method based on pyramid pooling module | |
| CN105631426B (en) | The method and device of text detection is carried out to picture | |
| CN107506792B (en) | Semi-supervised salient object detection method | |
| CN105096326A (en) | Laplace cutout matrix method by using moving least square method | |
| CN109559230B (en) | Bank transaction group discovery method and system based on overlapping community discovery algorithm | |
| CN104732545A (en) | Texture image segmentation method combined with sparse neighbor propagation and rapid spectral clustering | |
| CN108416343A (en) | A kind of facial image recognition method and device | |
| CN112270332A (en) | Three-dimensional target detection method and system based on sub-stream sparse convolution | |
| CN104077765B (en) | Image segmentation device, image partition method | |
| CN112184547B (en) | Super resolution method of infrared image and computer readable storage medium | |
| CN110555461A (en) | scene classification method and system based on multi-structure convolutional neural network feature fusion | |
| CN111160229A (en) | Video target detection method and device based on SSD (solid State disk) network | |
| CN109711269A (en) | A kind of classification hyperspectral imagery algorithm for composing empty Fusion Features based on 3D convolution | |
| CN109389553A (en) | Meteorological causes isopleth interpolation method based on T batten | |
| CN116664419A (en) | InSAR phase unwrapping method of multi-scale feature fusion noise reduction CNN network | |
| CN105551005A (en) | Quick image restoration method of total variation model coupled with gradient fidelity term | |
| CN111985488B (en) | Target detection segmentation method and system based on offline Gaussian model | |
| CN111144407A (en) | Target detection method, system, device and readable storage medium | |
| CN116977064A (en) | Wind control model construction method, system and device based on loss function | |
| CN111310772B (en) | Point line characteristic selection method and system for binocular vision SLAM | |
| CN110136164B (en) | Method for removing dynamic background based on online transmission transformation and low-rank sparse matrix decomposition | |
| CN115995024A (en) | Image classification method based on class diagram neural network | |
| CN106530319A (en) | Video object cooperative segmentation method based on track directed graph | |
| CN110807462A (en) | Training method insensitive to context of semantic segmentation model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 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 |