CN110232699A - A kind of image multilayer feature decomposition method based on the sparse statistical property of L0 - Google Patents

A kind of image multilayer feature decomposition method based on the sparse statistical property of L0 Download PDF

Info

Publication number
CN110232699A
CN110232699A CN201910367409.XA CN201910367409A CN110232699A CN 110232699 A CN110232699 A CN 110232699A CN 201910367409 A CN201910367409 A CN 201910367409A CN 110232699 A CN110232699 A CN 110232699A
Authority
CN
China
Prior art keywords
image
decomposed
decomposition method
decomposition
texture
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN201910367409.XA
Other languages
Chinese (zh)
Inventor
陈华松
丁琴
冯前胜
强豪
范媛媛
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Huaiyin Institute of Technology
Original Assignee
Huaiyin Institute of Technology
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Huaiyin Institute of Technology filed Critical Huaiyin Institute of Technology
Priority to CN201910367409.XA priority Critical patent/CN110232699A/en
Publication of CN110232699A publication Critical patent/CN110232699A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/12—Edge-based segmentation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/181—Segmentation; Edge detection involving edge growing; involving edge linking

Abstract

The present invention relates to technical field of image processing, disclose a kind of image multilayer feature decomposition method based on the sparse statistical property of L0, the model of the decomposition method are as follows:Wherein, f is image to be decomposed, u1For image outline part to be decomposed, u2It is image texture part to be decomposed, D1It is Wavelet tight frames operator, D2It is discrete cosine transform, α1, α2,β1It is decomposition coefficient,It is resolution error control item.Compared with prior art, this method can not only preferably statistical picture profile and texture information sparse characteristic, and there is better image structural information discomposing effect.

Description

A kind of image multilayer feature decomposition method based on the sparse statistical property of L0
Technical field
The present invention relates to technical field of image processing, in particular to a kind of image multilayer based on the sparse statistical property of L0 is special Levy decomposition method.
Background technique
Currently, intelligence system in society using more and more extensive, and the information Perception system master that intelligence system relies on A part is wanted to be derived from image information, thus, how to extract the characteristic information of image object is that intelligence system is normally transported Capable key.It is shown according to newest image analysis result of study, includes (usual point of information characteristics of different scale in image It is as shown in Figure 1 for the profile information of large scale and the texture information of small scale).Every kind of information all has different geometrical characteristics, These features make correct response for intelligence system accurate judgement target and play an important role.The technology of picture breakdown at present It is mainly based upon partial differential equation and based on wavelet transformation, the method for partial differential mainly uses the L1 norm i.e. image of the gradient of image Total difference of gradient corresponding with its dual function difference raised profile and texture, this method are lacked due to the model of total difference It falls into, so that information distortion at the image sharp edges proposed；And the method based on sparse expression is to be based on L1 norm statistical property, Its sparse expression scarce capacity to image outline and texture information.
Summary of the invention
Goal of the invention: aiming at the problems existing in the prior art, the present invention provides a kind of based on the sparse statistical property of L0 Image multilayer feature decomposition method, this method can not only preferably statistical picture profile and texture information sparse characteristic, and With better image characteristic information discomposing effect.
Technical solution: the present invention provides a kind of image multilayer feature decomposition methods based on the sparse statistical property of L0, should The model of decomposition method are as follows:Wherein, f is image to be decomposed, u1For Image outline part to be decomposed, u2It is image texture part to be decomposed, D1It is Wavelet tight frames operator, D2It is discrete cosine Transformation, α1, α2,β1It is decomposition coefficient,It is resolution error control item.
Further, the specific implementation step of the image multilayer feature decomposition method based on the sparse statistical property of L0 It is as follows:
Step 1: reading image to be decomposed, and initial value: resolution error tol=10 is arranged-5With Breaking Recurrently number m, and set Outline portion initial value u1=0, texture part initial value u2=0；
Step 2:while (i < m and error < tol)
Step 3: output u1And u2。
The utility model has the advantages that present invention application L0 norm carries out the contour structure of image large scale and the texture structure of small scale Sparse regularization decomposes image different structure using variance minimum principle；This method can not only better describe figure It is compared as the sparse statistical property of different structure feature, and with existing picture breakdown method, the image decomposited has more Good contour structure and texture and structural characteristic discrimination.
Detailed description of the invention
Fig. 1 is the image of image object multiple features decomposition analysis；It (a) is image；It (b) is the outline portion of image；(c) it is The texture part of image；
Fig. 2 is that image multi-layer information decomposition result compares: (a) contour structure and line of the decomposition of total difference+Dual Method Structure is managed, (b) contour structure and texture structure that total difference+L1 norm is decomposed, (c) fast decoupled filter equalizer contour structure Contour structure and texture structure, (e) L1 norm sparse decomposition contour structure and line are decomposed with texture structure (d) directional filters Structure is managed, (f) contour structure and texture structure based on the sparse Statistics decomposition of L0 proposed；
Fig. 3 is image object regional area decomposition result: (a) the decomposition contour structure and texture of total difference+Dual Method Structure, (b) total difference+L1 norm decomposes contour structure and texture structure, (c) fast decoupled filter equalizer contour structure and line It manages structure (d) directional filters and decomposes contour structure and texture structure, (e) L1 norm sparse decomposition contour structure and texture knot Structure, (f) the decomposition contour structure and texture structure based on the sparse statistics of L0 proposed.
Specific embodiment
The present invention is described in detail with reference to the accompanying drawing.
Embodiment 1:
Present embodiments provide for a kind of image multilayer feature decomposition method based on the sparse statistical property of L0, the decomposition sides The model of method are as follows:Wherein, f is image to be decomposed, u1It is to be decomposed Image outline part, u2It is image texture part to be decomposed, D1It is Wavelet tight frames operator, D2It is discrete cosine transform, α1, α2,β1It is decomposition coefficient,It is resolution error control item.
The specific implementation step of above-mentioned decomposition method is as follows:
Step 1: reading image to be decomposed, and initial value: resolution error tol=10 is arranged-5With Decomposition iteration number m=500, And set outline portion initial value u1=0, texture part initial value u2=0.U=f；
Step 2: how the number of iterations is less than or equal to m
Step 3: output display contour structure image u1With texture structure image u2。
Wherein, λ1,λ2It is the iterative parameter of image spatial feature decomposable process, for controlling the speed of iteration.
Image multilayer feature is decomposed by above-mentioned algorithm, decomposition result is as shown in Figures 2 and 3.By this implementation Picture breakdown effect in mode is good, and the outline portion and texture part after decomposition have preferable discrimination, preferably embodies The structure Embarrassing for having gone out contour structure and difference on texture structure scale and having generated is anisotropic.Moreover, the image after present invention decomposition Contour structure and texture structure have better visual analysis.
The technical concepts and features of above embodiment only to illustrate the invention, its object is to allow be familiar with technique People cans understand the content of the present invention and implement it accordingly, and it is not intended to limit the scope of the present invention.It is all according to the present invention The equivalent transformation or modification that Spirit Essence is done, should be covered by the protection scope of the present invention.

Claims (2)

1. a kind of image multilayer feature decomposition method based on the sparse statistical property of L0, which is characterized in that the mould of the decomposition method Type are as follows:Wherein, f is image to be decomposed, u1For image to be decomposed Outline portion, u2It is image texture part to be decomposed, D1It is Wavelet tight frames operator, D2It is discrete cosine transform, α1, α2,β1 It is decomposition coefficient,It is resolution error control item.
2. the image multilayer feature decomposition method according to claim 1 based on the sparse statistical property of L0, which is characterized in that The specific implementation step of the decomposition method is as follows:
Step1: reading image to be decomposed, and initial value: resolution error tol=10 is arranged-5With Breaking Recurrently number m, and profile is set Part initial value u1=0, texture part initial value u2=0；
Step2:while (i < m and error < tol)
Circulation executes
Step3: output u1And u2。
CN201910367409.XA 2019-05-05 2019-05-05 A kind of image multilayer feature decomposition method based on the sparse statistical property of L0 Pending CN110232699A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910367409.XA CN110232699A (en) 2019-05-05 2019-05-05 A kind of image multilayer feature decomposition method based on the sparse statistical property of L0

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910367409.XA CN110232699A (en) 2019-05-05 2019-05-05 A kind of image multilayer feature decomposition method based on the sparse statistical property of L0

Publications (1)

Publication Number Publication Date
CN110232699A true CN110232699A (en) 2019-09-13

Family

ID=67860596

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910367409.XA Pending CN110232699A (en) 2019-05-05 2019-05-05 A kind of image multilayer feature decomposition method based on the sparse statistical property of L0

Country Status (1)

Country Link
CN (1) CN110232699A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111709962A (en) * 2020-05-28 2020-09-25 淮阴工学院 Image contour and texture feature decomposition method based on anisotropic L0 gradient sparse expression and DCT (discrete cosine transformation)
CN111754428A (en) * 2020-06-11 2020-10-09 淮阴工学院 Image enhancement method and system based on anisotropic gradient model

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20070071331A1 (en) * 2005-09-24 2007-03-29 Xiteng Liu Image compression by economical quaternary reaching method
CN104484884A (en) * 2014-12-30 2015-04-01 天津大学 Intrinsic image decomposition method based on multi-scale L0 sparse constraint
CN106023122A (en) * 2016-05-05 2016-10-12 江南大学 Image fusion method based on multi-channel decomposition
CN107492070A (en) * 2017-07-10 2017-12-19 华北电力大学 A kind of single image super-resolution computational methods of binary channels convolutional neural networks

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20070071331A1 (en) * 2005-09-24 2007-03-29 Xiteng Liu Image compression by economical quaternary reaching method
CN104484884A (en) * 2014-12-30 2015-04-01 天津大学 Intrinsic image decomposition method based on multi-scale L0 sparse constraint
CN106023122A (en) * 2016-05-05 2016-10-12 江南大学 Image fusion method based on multi-channel decomposition
CN107492070A (en) * 2017-07-10 2017-12-19 华北电力大学 A kind of single image super-resolution computational methods of binary channels convolutional neural networks

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
HUASONG CHEN 等: "A structural decomposition based method for restoring images corrupted by impulse noise", 《PROCEEDINGS OF SPIE》 *
陈华松: "基于各向异性总变分和稀疏分解的图像复原方法及其应用研究", 《中国博士学位论文全文数据库 信息科技辑》 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111709962A (en) * 2020-05-28 2020-09-25 淮阴工学院 Image contour and texture feature decomposition method based on anisotropic L0 gradient sparse expression and DCT (discrete cosine transformation)
CN111754428A (en) * 2020-06-11 2020-10-09 淮阴工学院 Image enhancement method and system based on anisotropic gradient model

Similar Documents

Publication Publication Date Title
CN108230264B (en) Single image defogging method based on ResNet neural network
CN110428433B (en) Canny edge detection algorithm based on local threshold
CN112424828B (en) Nuclear fuzzy C-means quick clustering algorithm integrating space constraint
CN109767439B (en) Target detection method for multi-scale difference and bilateral filtering of self-adaptive window
CN108564549B (en) Image defogging method based on multi-scale dense connection network
CN106127688B (en) A kind of super-resolution image reconstruction method and its system
CN109146797B (en) A kind of impulsive noise ancient book image inpainting method sparse based on Lp pseudonorm and overlapping group
CN105184744A (en) Prior fuzzy kernel estimation method based on standardized sparse measurement image block
CN110232699A (en) A kind of image multilayer feature decomposition method based on the sparse statistical property of L0
CN107507222B (en) Anti-occlusion particle filter target tracking method based on integral histogram
CN108734177B (en) Double-step correlation filtering target tracking method
CN105701785A (en) Image smog removing method based on sky region division and transmissivity optimization of weighting TV
CN102170526A (en) Method for calculation of defocus fuzzy core and sharp processing of defocus fuzzy image of defocus fuzzy core
CN102135606A (en) KNN (K-Nearest Neighbor) sorting algorithm based method for correcting and segmenting grayscale nonuniformity of MR (Magnetic Resonance) image
CN107220991A (en) A kind of Robust Real-time Moving Object Tracking based on compressed sensing
CN110751667B (en) Method for detecting infrared dim and small targets under complex background based on human visual system
CN115689932A (en) Image defogging method based on deep neural network
CN112907461B (en) Defogging enhancement method for infrared foggy-day degraded image
CN109859131A (en) A kind of image recovery method based on multi-scale self-similarity Yu conformal constraint
CN106023097A (en) Iterative-method-based flow field image preprocessing algorithm
CN105303538A (en) Gauss noise variance estimation method based on NSCT and PCA
CN110880192B (en) Image DCT coefficient distribution fitting method based on probability density function dictionary
CN104616266B (en) A kind of noise variance estimation method based on broad sense autoregression heteroscedastic model
CN103955893A (en) Image denoising method based on separable total variation model
CN116363168A (en) Remote sensing video target tracking method and system based on super-resolution network

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20190913