CN110400300A - Lesion vessels accurate detecting method based on Block- matching adaptive weighting rarefaction representation - Google Patents
Lesion vessels accurate detecting method based on Block- matching adaptive weighting rarefaction representation Download PDFInfo
- Publication number
- CN110400300A CN110400300A CN201910669052.0A CN201910669052A CN110400300A CN 110400300 A CN110400300 A CN 110400300A CN 201910669052 A CN201910669052 A CN 201910669052A CN 110400300 A CN110400300 A CN 110400300A
- Authority
- CN
- China
- Prior art keywords
- block
- rarefaction representation
- information
- blood vessel
- parameter
- 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
- 238000000034 method Methods 0.000 title claims abstract description 37
- 230000003902 lesion Effects 0.000 title claims abstract description 35
- 230000003044 adaptive effect Effects 0.000 title claims abstract description 21
- 210000004204 blood vessel Anatomy 0.000 claims abstract description 37
- 238000012549 training Methods 0.000 claims abstract description 24
- 239000000284 extract Substances 0.000 claims abstract description 6
- 238000005457 optimization Methods 0.000 claims description 33
- 239000011159 matrix material Substances 0.000 claims description 14
- 239000013598 vector Substances 0.000 claims description 9
- 239000008280 blood Substances 0.000 claims description 7
- 210000004369 blood Anatomy 0.000 claims description 7
- 230000011218 segmentation Effects 0.000 claims description 7
- 238000013507 mapping Methods 0.000 claims description 6
- 238000010276 construction Methods 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 3
- 238000012163 sequencing technique Methods 0.000 claims description 2
- 230000001360 synchronised effect Effects 0.000 claims description 2
- 239000004576 sand Substances 0.000 claims 1
- 238000001514 detection method Methods 0.000 abstract description 13
- 238000004364 calculation method Methods 0.000 abstract description 4
- 230000009286 beneficial effect Effects 0.000 abstract description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 241000208340 Araliaceae Species 0.000 description 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 1
- 235000003140 Panax quinquefolius Nutrition 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 235000008434 ginseng Nutrition 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/20—Image enhancement or restoration using local operators
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/68—Analysis of geometric attributes of symmetry
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20024—Filtering details
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30101—Blood vessel; Artery; Vein; Vascular
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Geometry (AREA)
- Data Mining & Analysis (AREA)
- Radiology & Medical Imaging (AREA)
- Artificial Intelligence (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- General Health & Medical Sciences (AREA)
- Quality & Reliability (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Medical Informatics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses the lesion vessels accurate detecting methods based on Block- matching adaptive weighting rarefaction representation, generate multiple dimensioned blood vessel training set；Block of information based on vessels axis extracts；Multi-scale dictionary study based on rarefaction representation adaptively captures local patholoic change blood vessel feature by rarefaction representation algorithm from block message library；Lesion vessels are carried out using adaptive weighting rarefaction representation sorting algorithm accurately to detect.The beneficial effects of the present invention are: the present invention is directed to the difficult point that lesion vessels can not be detected precisely, it proposes to be based on Block- matching adaptive weighting rarefaction representation detection algorithm, the detection accuracy of lesion vessels is greatly improved using redundancy included in vessel information block, and reduces the calculation amount of rarefaction representation detection algorithm using adaptive weighting algorithm and multiple dimensioned training set.
Description
Technical field
The invention belongs to medicine technology field, it is accurate to be related to the lesion vessels based on Block- matching adaptive weighting rarefaction representation
Detection method.
Background technique
Currently, blood vessel segmentation technology achieves higher segmentation precision for the healthy blood vessel under complex environment, but face
It will appear higher false detection rate and omission factor when to lesion vessels, so that testing result lacks Clinical significance of MG.It is contained in more
Redundancy in scale training set is the key that adaptive weighting sparse representation method realizes that lesion vessels precisely detect.So
And huge calculation amount brought by sparse representation method seriously reduces the convergence of detection algorithm.
Summary of the invention
The purpose of the present invention is to provide the accurate sides of detection of the lesion vessels based on Block- matching adaptive weighting rarefaction representation
Method, the beneficial effects of the present invention are: being proposed based on Block- matching certainly for the difficult point that lesion vessels in CT image can not be detected precisely
Weight rarefaction representation detection algorithm is adapted to, greatly improves lesion vessels using redundancy included in vessel information block
Detection accuracy, and reduce using adaptive weighting algorithm and multiple dimensioned training set the calculation amount of rarefaction representation detection algorithm.
The technical scheme adopted by the invention is that following the steps below:
Step 1: generating multiple dimensioned blood vessel training set；
Step 2: the block of information based on vessels axis extracts；
Step 3: the multi-scale dictionary study based on rarefaction representation, by rarefaction representation algorithm, adaptively from block message
Local patholoic change blood vessel feature is captured in library；
Step 4: lesion vessels being carried out using adaptive weighting rarefaction representation sorting algorithm and are accurately detected.
Further, step 1 is that blood vessel structure is regarded as to the intracorporal linear structure of people, using Hessian matrix construction based on height
The multidimensional linear structure filter of this convolution is more using what is constructed on the basis of the blood vessel training set of traditional semi-automatic segmentation
Linear structure filter is tieed up, blood vessel training set is converted into training set relevant to blood vessel scale.
Further, step 2 is to extract the central axis of lesion vessels, edge using multiple dimensioned sublevel layer vessel tracking algorithm
Axis cross-wise direction from multiple dimensioned blood vessel training set, extracts vessel information block, sets up more using blood vessel out to out as radius
The size in scale block message library, block of information is 7 × 7 × 7 voxels, and scale is determined by the target blood scale extracted.
Further, step 3 process is as follows:
Wherein dsThe block message library for being s for scale, β are the coefficient vector of rarefaction representation, and p is the block of information in training set, ds
It is provided with the Optimization Solution of β by rarefaction representation algorithm, the residual error of rarefaction representation determines the representational of block of information, and representational high
Block of information will be placed into multi-scale dictionary, realize the capture of lesion vessels local feature.Multi-scale dictionary learns (step 3) and dilute
Dredging indicates that Optimization Solution (step 4) iteration carries out.A threshold residual value is set, when the rarefaction representation residual error of block in information bank is greater than
When threshold value, which will be placed into dictionary, realize the renewal learning of multi-scale dictionary.
Further, step 4 process is as follows:
Wherein p is target blood block of information to be detected；Representing matrix element multiplication；W indicates block message library dsFrom
Adapt to weight vectors；W is the relevant mapping matrix of weight；aiAnd ajFor block message library dsMiddle any two block of information；λ1、λ2And λ3
For non-negative influence factor parameter；For the normalization of mapping matrix W, block message library d can reducesIn similar letter
Cease the characteristic distance of block；According to the rarefaction representation residual error of target information block p, the similarity of p and blood vessel, following formula institute are defined
Show:
Wherein β is that rarefaction representation coefficient obtained by formula (2) optimize is recognized when target information block and higher blood vessel similarity
The block of information that sets the goal p is a part of blood vessel, realizes the classification of target information block；The collection process of target information block p to be detected
It is consistent with the collection process of training set block of information, i.e., along vessels axis cross-wise direction.Therefore, target information block constitute one it is to be checked
Sequencing column { p0,...,pn-1,pn,...}.Due to closing on information block position to be detected, the similitude between proximity information block makes
pn-1Rarefaction representation parameter can guide pnParameter optimization, therefore, the Optimization Solution process of formula (2) is divided into following five step: 1)
In pn-1Scale sn-1Search p nearbynOptimal scale sn；2) in parameter wn-1、Wn-1And snIn the case where determination, Optimization Solution ginseng
Number βn, in this case, formula (2) is converted into simple traditional rarefaction representation Optimization Solution problem, can be readily available optimal
Solution:WhereinG is about XnDiagonal matrix；3) joining
Number Wn-1、βnAnd snIn the case where determination, Optimization Solution parameter wn, in this case, the Lagrangian of formula (2) meets
Karush-Kuhn-Tucker synchronizes the condition of convergence, and then obtains wnOptimal solution:Whereinεm-l+1For the m-l+1 element of ε, m βnIn element number, l βnMiddle nonzero element
Number；4) in parameter wn、βnAnd snIn the case where determination, Optimization Solution parameter Wn, in this case, formula (2) becomes without constraint most
Optimization problem can be used traditional CG Schmidt (2005) Optimizing operator and be solved；5) in parameter wn、WnAnd snIt determines
In the case where, the Optimization Solution process of second step is repeated, undated parameter β is optimizedn。
Specific embodiment
The present invention is described in detail With reference to embodiment.
The present invention is based on the lesion vessels accurate detecting method of Block- matching adaptive weighting rarefaction representation, steps are as follows:
Step 1: generating multiple dimensioned blood vessel training set.Regard blood vessel structure as people intracorporal linear structure, utilizes Hai Sen
Multidimensional linear structure filter of the matrix construction based on Gaussian convolution, the wherein standard deviation parameter of filter and blood vessel scale phase
It closes.On the basis of the blood vessel training set of traditional semi-automatic segmentation, using the multidimensional linear structure filter constructed, blood vessel is instructed
Practice collection and is converted into training set relevant to blood vessel scale.
Step 2: the block of information based on vessels axis extracts.Using multiple dimensioned sublevel layer vessel tracking algorithm, lesion is extracted
The central axis of blood vessel.Along axis cross-wise direction, using blood vessel out to out as radius, vessel information block is extracted, is set up multiple dimensioned
Block message library.The size of block of information is 7 × 7 × 7 voxels, and scale is determined by the target blood scale extracted.
Multiple dimensioned block message library can capture that lesion vessels are changeable, the lower local feature of robustness, realize detection essence
The unification of degree and Algorithms T-cbmplexity.
Step 3: the multi-scale dictionary study based on rarefaction representation.By rarefaction representation algorithm, adaptively from block message
Local patholoic change blood vessel feature is captured in library, detailed process is as follows:
Wherein dsThe block message library for being s for scale, β are the coefficient vector of rarefaction representation, and p is the block of information in training set.ds
It is provided with the Optimization Solution of β by rarefaction representation algorithm, the residual error of rarefaction representation determines the representational of block of information, and representational high
Block of information will be placed into multi-scale dictionary, realize the capture of lesion vessels local feature.Multi-scale dictionary learns (step 3) and dilute
Dredging indicates that Optimization Solution (step 4) iteration carries out.A threshold residual value is set, when the rarefaction representation residual error of block in information bank is greater than
When threshold value, which will be placed into dictionary, realize the renewal learning of multi-scale dictionary.
Step 4: lesion vessels being carried out using adaptive weighting rarefaction representation sorting algorithm and are accurately detected.Traditional rarefaction representation
Algorithm is by minimizing residual errorFind target blood rarefaction representation vector, and lesion vessels it is changeable,
The lower local feature of robustness brings huge calculation amount to the minimum of residual error, so that traditional rarefaction representation algorithm is being examined
It can not be restrained when surveying lesion vessels.For this difficult point, the present invention is relevant certainly by adding characteristic distance to each block of information
Weight is adapted to, the Optimization Solution process of rarefaction representation vector is guided, so that the use of multi-scale dictionary is more efficient, and then is realized
The accurate detection of lesion vessels.The process of adaptive weighting rarefaction representation is as follows:
Wherein p is target blood block of information to be detected；Representing matrix element multiplication；W indicates block message library dsFrom
Adapt to weight vectors；W is the relevant mapping matrix of weight；aiAnd ajFor block message library dsMiddle any two block of information；λ1、λ2And λ3
For non-negative influence factor parameter；For the normalization of mapping matrix W, block message library d can reducesIn similar letter
Cease the characteristic distance of block.
According to the rarefaction representation residual error of target information block p, the similarity of p and blood vessel are defined, shown in following formula:
Wherein β is that formula (2) optimize gained rarefaction representation coefficient.When target information block and higher blood vessel similarity, recognize
The block of information that sets the goal p is a part of blood vessel, realizes the classification of target information block.So far, the test problems of lesion vessels sum up
To the Optimization Solution problem of formula (2).
The collection process of target information block p to be detected is consistent with the collection process of training set block of information, i.e., along vessel axis
Line cross-wise direction, therefore, target information block constitute a sequence { p to be detected0,...,pn-1,pn,...}.Due to letter to be detected
Closing on breath block position, the similitude between proximity information block makes pn-1Rarefaction representation parameter can guide pnParameter optimization.
Therefore, the Optimization Solution process of formula (2) is divided into following five step: 1) in pn-1Scale sn-1Search p nearbynOptimal scale sn；2)
In parameter wn-1、Wn-1And snIn the case where determination, Optimization Solution parameter betan.In this case, formula (2) is converted into simple tradition
Rarefaction representation Optimization Solution problem, can be readily available optimal solution:Its
InG is about XnDiagonal matrix；3) in parameter Wn-1、βnAnd snIn the case where determination, Optimization Solution parameter
wn.In this case, the Lagrangian of formula (2) meets the synchronous condition of convergence of Karush-Kuhn-Tucker, and then obtains
Obtain wnOptimal solution:
Whereinεm-l+1For the m-l+1 member of ε
Element, m βnIn element number, l βnMiddle nonzero element number；4) in parameter wn、βnAnd snIn the case where determination, Optimization Solution
Parameter Wn.In this case, formula (2) becomes Unconstrained Optimization Problem, and traditional CG Schmidt (2005) can be used
Optimizing operator is solved；5) in parameter wn、WnAnd snIn the case where determination, the Optimization Solution process of second step is repeated, optimization is more
New parameter βn。
By adding the relevant adaptive weighting of characteristic distance to each block of information in library, guidance rarefaction representation vector
Optimization Solution process so that the use of multi-scale dictionary is more efficient, and then realizes the accurate detection of lesion vessels.
The method of the present invention uses vessels axis guided constraint segmentation strategy, and uses multiple dimensioned training in the training process
Collection takes into account time complexity while realizing precisely segmentation.The experimental results showed that point of the method for the present invention for lesion vessels
It cuts accuracy rate and is up to 91%, be a kind of precisely efficient lesion vessels dividing method.
The above is only not to make limit in any form to the present invention to better embodiment of the invention
System, any simple modification that embodiment of above is made according to the technical essence of the invention, equivalent variations and modification,
Belong in the range of technical solution of the present invention.
Claims (5)
1. the lesion vessels accurate detecting method based on Block- matching adaptive weighting rarefaction representation, it is characterised in that according to following step
It is rapid to carry out:
Step 1: generating multiple dimensioned blood vessel training set；
Step 2: the block of information based on vessels axis extracts；
Step 3: the multi-scale dictionary study based on rarefaction representation, through rarefaction representation algorithm, adaptively from block message library
Capture local patholoic change blood vessel feature；
Step 4: lesion vessels being carried out using adaptive weighting rarefaction representation sorting algorithm and are accurately detected.
2. according to the lesion vessels accurate detecting method based on Block- matching adaptive weighting rarefaction representation described in claim 1,
Be characterized in that: the step 1 is that blood vessel structure is regarded as to the intracorporal linear structure of people, is based on Gauss using Hessian matrix construction
The multidimensional linear structure filter of convolution, on the basis of the blood vessel training set of traditional semi-automatic segmentation, using the multidimensional constructed
Blood vessel training set is converted into training set relevant to blood vessel scale by linear structure filter.
3. according to the lesion vessels accurate detecting method based on Block- matching adaptive weighting rarefaction representation described in claim 1,
Be characterized in that: the step 2 is the central axis of lesion vessels to be extracted, along axis using multiple dimensioned sublevel layer vessel tracking algorithm
Line cross-wise direction extracts vessel information block, sets up multiple dimensioned block message library using blood vessel out to out as radius, block of information it is big
Small is 7 × 7 × 7 voxels, and scale is determined by the target blood scale extracted.
4. according to the lesion vessels accurate detecting method based on Block- matching adaptive weighting rarefaction representation described in claim 1,
Be characterized in that: step 3 process is as follows:
Wherein dsThe block message library for being s for scale, β are the coefficient vector of rarefaction representation, and p is the block of information in training set, dsAnd β
Optimization Solution provided by rarefaction representation algorithm, the residual error of rarefaction representation determines the representational of block of information, and representational high letter
Breath block will be placed into multi-scale dictionary, realize the capture of lesion vessels local feature.
5. according to the lesion vessels accurate detecting method based on Block- matching adaptive weighting rarefaction representation described in claim 1,
Be characterized in that: step 4 process is as follows:
Wherein p is target blood block of information to be detected；Representing matrix element multiplication；W indicates block message library dsAdaptive power
Weight vector；W is the relevant mapping matrix of weight；aiAnd ajFor block message library dsMiddle any two block of information；λ1、λ2And λ3It is non-negative
Influence factor parameter；For the normalization of mapping matrix W, block message library d can reducesMiddle analog information block
Characteristic distance；According to the rarefaction representation residual error of target information block p, the similarity of p and blood vessel are defined, shown in following formula:
Wherein β is that formula (2) optimize gained rarefaction representation coefficient, when target information block and higher blood vessel similarity, assert mesh
A part that block of information p is blood vessel is marked, realizes the classification of target information block；The collection process of target information block p to be detected and
The collection process of training set block of information is consistent, i.e., along vessels axis cross-wise direction, therefore, target information block constitute one it is to be checked
Sequencing column { p0,...,pn-1,pn... }, due to closing on information block position to be detected, the similitude between proximity information block makes
Obtain pn-1Rarefaction representation parameter can guide pnParameter optimization, therefore, the Optimization Solution process of formula (2) is divided into following five
Step: 1) in pn-1Scale sn-1Search p nearbynOptimal scale sn；2) in parameter wn-1、Wn-1And snIn the case where determination, optimization
Solve parameter betan, in this case, formula (2) is converted into simple traditional rarefaction representation Optimization Solution problem, can be easy to
Obtain optimal solution:WhereinG is about XnTo angular moment
Battle array；3) in parameter Wn-1、βnAnd snIn the case where determination, Optimization Solution parameter wn, in this case, the Lagrangian letter of formula (2)
Number meets the synchronous condition of convergence of Karush-Kuhn-Tucker, and then obtains wnOptimal solution:
Whereinεm-l+1For the m-l+1 element of ε, m βnIn element number, l βnMiddle non-zero entry
Plain number；4) in parameter wn、βnAnd snIn the case where determination, Optimization Solution parameter Wn, in this case, formula (2) becomes without about
Beam optimization problem can be used traditional CGSchmidt (2005) Optimizing operator and be solved；5) in parameter wn、WnAnd snReally
In the case where fixed, the Optimization Solution process of second step is repeated, optimizes undated parameter βn。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910669052.0A CN110400300B (en) | 2019-07-24 | 2019-07-24 | Pathological blood vessel accurate detection method based on block matching self-adaptive weight sparse representation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910669052.0A CN110400300B (en) | 2019-07-24 | 2019-07-24 | Pathological blood vessel accurate detection method based on block matching self-adaptive weight sparse representation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110400300A true CN110400300A (en) | 2019-11-01 |
| CN110400300B CN110400300B (en) | 2023-06-23 |
Family
ID=68325802
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910669052.0A Active CN110400300B (en) | 2019-07-24 | 2019-07-24 | Pathological blood vessel accurate detection method based on block matching self-adaptive weight sparse representation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110400300B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112184840A (en) * | 2020-09-22 | 2021-01-05 | 上海交通大学 | 3D point cloud compression system based on multi-scale structured dictionary learning |
| LU102391B1 (en) * | 2021-01-07 | 2022-04-28 | Harbin Inst Technology Weihai | Method for accurate detection of blood vessels with pathological changes based on adaptive weighted sparse representation of block matching |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080069445A1 (en) * | 2003-03-07 | 2008-03-20 | Martin Weber | Image processing apparatus and methods |
| US20150287223A1 (en) * | 2014-04-04 | 2015-10-08 | The Board Of Trustees Of The University Of Illinois | Highly accelerated imaging and image reconstruction using adaptive sparsifying transforms |
| CN105678343A (en) * | 2015-08-25 | 2016-06-15 | 浙江工业大学 | Adaptive-weighted-group-sparse-representation-based diagnosis method for noise abnormity of hydroelectric generating set |
| US20160284085A1 (en) * | 2015-03-25 | 2016-09-29 | Oregon Health & Science University | Systems and methods of choroidal neovascularization detection using optical coherence tomography angiography |
| CN109492570A (en) * | 2018-11-02 | 2019-03-19 | 电子科技大学 | A kind of SAR image target recognition method based on multiple dimensioned rarefaction representation |
-
2019
- 2019-07-24 CN CN201910669052.0A patent/CN110400300B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080069445A1 (en) * | 2003-03-07 | 2008-03-20 | Martin Weber | Image processing apparatus and methods |
| US20150287223A1 (en) * | 2014-04-04 | 2015-10-08 | The Board Of Trustees Of The University Of Illinois | Highly accelerated imaging and image reconstruction using adaptive sparsifying transforms |
| US20160284085A1 (en) * | 2015-03-25 | 2016-09-29 | Oregon Health & Science University | Systems and methods of choroidal neovascularization detection using optical coherence tomography angiography |
| CN105678343A (en) * | 2015-08-25 | 2016-06-15 | 浙江工业大学 | Adaptive-weighted-group-sparse-representation-based diagnosis method for noise abnormity of hydroelectric generating set |
| CN109492570A (en) * | 2018-11-02 | 2019-03-19 | 电子科技大学 | A kind of SAR image target recognition method based on multiple dimensioned rarefaction representation |
Non-Patent Citations (2)
| Title |
|---|
| 王阳;夏利民;王嘉;: "基于加权分块稀疏表示的步态识别" * |
| 薛文渲;刘建霞;刘然;袁晓辉;: "改进U型网络的眼底视网膜血管分割方法" * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112184840A (en) * | 2020-09-22 | 2021-01-05 | 上海交通大学 | 3D point cloud compression system based on multi-scale structured dictionary learning |
| US11836954B2 (en) | 2020-09-22 | 2023-12-05 | Shanghai Jiao Tong University | 3D point cloud compression system based on multi-scale structured dictionary learning |
| LU102391B1 (en) * | 2021-01-07 | 2022-04-28 | Harbin Inst Technology Weihai | Method for accurate detection of blood vessels with pathological changes based on adaptive weighted sparse representation of block matching |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110400300B (en) | 2023-06-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108052896B (en) | Human body behavior identification method based on convolutional neural network and support vector machine | |
| CN107145842B (en) | Face recognition method combining LBP characteristic graph and convolutional neural network | |
| Hu et al. | Gait-based gender classification using mixed conditional random field | |
| CN106407958B (en) | Face feature detection method based on double-layer cascade | |
| CN104866829A (en) | Cross-age face verify method based on characteristic learning | |
| CN104881671B (en) | A kind of high score remote sensing image Local Feature Extraction based on 2D Gabor | |
| CN109584283A (en) | A kind of Medical Image Registration Algorithm based on convolutional neural networks | |
| CN111199207B (en) | Two-dimensional multi-human body posture estimation method based on depth residual error neural network | |
| CN106548194B (en) | The construction method and localization method of two dimensional image human joint points location model | |
| CN109697692A (en) | One kind being based on the similar feature matching method of partial structurtes | |
| Gupta et al. | Lung cancer detection using curvelet transform and neural network | |
| CN112598613A (en) | Determination method based on depth image segmentation and recognition for intelligent lung cancer diagnosis | |
| Guo et al. | Msanet: multiscale aggregation network integrating spatial and channel information for lung nodule detection | |
| CN107729926A (en) | A kind of data amplification method based on higher dimensional space conversion, mechanical recognition system | |
| CN110400300A (en) | Lesion vessels accurate detecting method based on Block- matching adaptive weighting rarefaction representation | |
| CN110334656A (en) | Multi-source Remote Sensing Images Clean water withdraw method and device based on information source probability weight | |
| CN108230313A (en) | Based on the adaptively selected SAR image object detection method with discrimination model of component | |
| Cheng et al. | DRNet: Towards fast, accurate and practical dish recognition | |
| CN106971176A (en) | Tracking infrared human body target method based on rarefaction representation | |
| Xu et al. | Application of artificial intelligence technology in medical imaging | |
| CN113516643A (en) | Method for detecting retinal vessel bifurcation and intersection points in OCTA image | |
| Yan et al. | Siamese-Prototypical Network with Data Augmentation Pre-training for Few-shot Medical Image Classification | |
| Hossain et al. | Streamlining Brain Tumor Classification with Custom Transfer Learning in MRI Images | |
| Huang et al. | Online action tube detection via resolving the spatio-temporal context pattern | |
| CN112699954A (en) | Closed-loop detection method based on deep learning and bag-of-words 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 |