CN105205810B - MR image hippocampus dividing method based on distance field fusion - Google Patents

MR image hippocampus dividing method based on distance field fusion Download PDF

Info

Publication number
CN105205810B
CN105205810B CN201510539859.4A CN201510539859A CN105205810B CN 105205810 B CN105205810 B CN 105205810B CN 201510539859 A CN201510539859 A CN 201510539859A CN 105205810 B CN105205810 B CN 105205810B
Authority
CN
China
Prior art keywords
image
training set
point
distance field
test
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.)
Active
Application number
CN201510539859.4A
Other languages
Chinese (zh)
Other versions
CN105205810A (en
Inventor
冯前进
庞树茂
阳维
卢振泰
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.)
Southern Medical University
Original Assignee
Southern Medical University
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 Southern Medical University filed Critical Southern Medical University
Priority to CN201510539859.4A priority Critical patent/CN105205810B/en
Publication of CN105205810A publication Critical patent/CN105205810A/en
Application granted granted Critical
Publication of CN105205810B publication Critical patent/CN105205810B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • 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
    • G06T2207/10—Image acquisition modality
    • G06T2207/10072—Tomographic images
    • G06T2207/10088—Magnetic resonance imaging [MRI]
    • 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
    • G06T2207/20021—Dividing image into blocks, subimages or windows
    • 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
    • G06T2207/30016—Brain

Landscapes

  • Magnetic Resonance Imaging Apparatus (AREA)

Abstract

A kind of MR image hippocampus dividing method based on distance field fusion.Include: (1) to initial MR test image normalized to be split, removes skull and biased field, the MR test image after obtaining normalized；(2) MR training set image and training set label image are registrated to MR test image；(3) the label image after registration is obtained into distance field DF by range conversion；(4) to the point in MR test imageTake image blockAnd it is converted to a column vector；(5) it defines search window and chooses MR dictionary and DF dictionary；(6) it usesLocal linear expresses test sample, seek dictionary weight coefficient vector；(7) test sample is obtainedDF forecast image block vector, and handleIt is converted into image block；(8) step (4)-(7) are repeated, the DF value of each point is obtained；(9) label corresponding to each point of test image is obtained.The method of the present invention can from MR image accurately segmentation object.

Description

MR image hippocampus dividing method based on distance field fusion
Technical field
The present invention relates to medical image analysis technical fields, and in particular to a kind of MR image hippocampus based on distance field fusion Body dividing method.
Background technique
Hippocampus belongs to the grey matter structure of midbrain, and hippocampus atrophy is Alzheimer's disease, schizophrenia, depression etc. A kind of pathological manifestations of mental disease can clinically make these mental diseases by the volume and form for analyzing hippocampus Diagnosis.MRI can provide high-resolution anatomical structure, clear contrast and extensive image sequence, this is facing MRI technique Bed diagnosis aspect is very popular.Doctor can analyze all brain structures of patient by the brain MR image of patient, thus Diagnosis is made to disease.The hippocampus of MR brain image is split, is measurement, the morphological analysis of hippocampus volume size Premise.Since the manual segmentation of hippocampus is very time-consuming, uninteresting, and the segmentation result of different doctor also can difference, sea Being segmented in clinical application automatically for horse body is particularly important.
Since the volume very little of hippocampus, complex-shaped, obscurity boundary, partial volume effect are obvious in MR image, hippocampus The automatic segmentation of body is very challenging.
So far, MR image hippocampus dividing method mainly has following two class:
The first kind is the method based on movable contour model (ACM).Such methods are according to image grayscale demographic information Or image gradient develops to curve (curved surface), final curves (curved surface) converge at object boundary.ACM method is to image Noise is more sensitive.
Second class is the method based on multichannel chromatogram.Such methods are atlas registration to test image, to the training after deformation Collect label image, label fusion is realized by some way, obtains final segmentation result.Label fusion do not use to The shape priors of segmentation object, segmentation precision are limited.
In view of the deficiencies of the prior art, the method for the present invention proposes a kind of MR image hippocampus segmentation based on distance field fusion Method.
Summary of the invention
The method of the present invention provides a kind of MR image hippocampus dividing method based on distance field fusion, the method for the present invention success Apply the hippocampus segmentation in MR brain image, the hippocampus in the accurate Ground Split MR image of energy in ground.
Above-mentioned purpose of the invention is realized by following technological means.
A kind of MR image hippocampus dividing method based on distance field fusion, this method are based on two kinds of hypothesis:
I, MR image block and DF image block are located on two non-linearity manifolds, any one MR image block can be where it Neighbour's sample in the local space of manifold is linearly expressed；
II, under local constraint, the mapping of MR manifold to DF manifold is similar to a differomorphism mapping；
The MR image hippocampus dividing method based on distance field fusion carries out as follows:
(1) to initial MR test image normalized to be split, skull and biased field is removed, normalized is obtained MR test image afterwards；
(2) by the MR training set image and training set label corresponding with MR training set image in pre-prepd training set Image registration is to MR test image, so that MR training set image, training set label image and MR test image are on spatial position Alignment；
(3) range conversion is carried out to the training set label image after step (2) registration, obtains the DF of training set label image Training set distance field, the value of DF training set distance field corresponding to point x are as follows:
Wherein C indicates the boundary of segmentation object, and B indicates the nearest point of distance x, and B ∈ C, dist (x, B) indicate point x Euclidean distance between point B；
(4) to the point x in MR test image, an image block x is taken centered on xMR, xMRIt is converted to a column vectorIt is used as the feature of point x, obtains test sampleWherein m representation vectorDimension；
(5) search window is defined centered on point x respectively on training set MR image, training set distance field DF image to choose Image blockWithConstruct MR dictionary With DF dictionaryN is the size of dictionary；
(6) based on the assumption that I, with MR dictionary DMRLocal linear indicates MR test sampleDictionary DMRWeight coefficient vectorIt is solved with LAE method, expression is as follows:
Indicate test sampleIn dictionary DMRIn k neighbour；
(7) step (6) resulting dictionary D is usedMRWeight coefficient vectorLinear combination DF dictionary xDFIn sample, obtain Test sampleDF forecast image block vectorAnd handleIt is converted into image block xDF；Specifically:
Based on the assumption that I, it can obtain:
Based on the assumption that II, it can obtain:
Since f is local linear, therefore:
?It is converted into image block xDF；
(8) each point in MR test image is repeated to operate according to step (4)-(7), obtains the DF training set of each point The value of distance field；
It is indicated centered on point x, size and x with P (x)DFEqual image block, for any point u in P (x), weight Are as follows:
The value of the DF training set distance field of point x are as follows:
Indicate the value of the DF training set distance field for the point x that the image block centered on point u is predicted；
(9) to the DF forecast image progress threshold process come is predicted, mark corresponding to each point of test image is obtained Number, specifically:
The distance field as defined in formula (1) is it is found that the label of point x can be exported by formula (8):
It indicates that the point belongs to segmentation object, and background is indicated marked as 0 marked as 1, hippocampus point is obtained according to segmentation object Cut image.
Preferably, above-mentioned steps (1) are specifically that MR test image is normalized using gray scale normalization method, Skull is removed with BET algorithm, removes biased field with N4 algorithm.
Preferably, above-mentioned steps (2) be specifically using DRAMMS software tool by training set MR training set image and Training set label image is registrated to MR test image.
Preferably, above-mentioned steps (4) are specifically to extract the feature of each point in MR test image, are used as test specimens This.
Preferably, above-mentioned steps (5) are specifically that local search is used in MR training set image and DF training set distance field Window constructs MR dictionary and DF dictionary.
Preferably, above-mentioned steps (8) are specifically to handle the DF forecast image block weighted average of overlapping, final to obtain each The predicted value of the DF training set distance field of point.
The method of the present invention provides a kind of MR image hippocampus dividing method based on distance field fusion, and the method for the present invention can be quasi- Hippocampus in true Ground Split MR image, can be used for the diagnosis of the mental diseases such as Alzheimer's disease.
Detailed description of the invention
Using attached drawing, the present invention is further illustrated, but the content in attached drawing is not constituted to any limit of the invention System.
Fig. 1 is the flow diagram of the method for the present invention；
Fig. 2 is the frame diagram of the method for the present invention；
Fig. 3 is the details flow diagram of the method for the present invention；
Fig. 4 is for the method for the present invention with different size of search window to the segmentation result box traction substation of 15 test images；Wherein, Fig. 4 (a) is the segmentation result box traction substation for right hippocampus, and Fig. 4 (b) is the segmentation result box traction substation for left hippocampus；
Fig. 5 is the segmentation result comparison diagram with label fusion method and distance field fusion method to 15 test images, In, Fig. 5 (a) is the segmentation result comparison diagram for right hippocampus, and Fig. 5 (b) is the segmentation result comparison diagram for left hippocampus；
Fig. 6 is the coronal-plane of the segmentation result in four test images using distance field fusion method and label fusion method Schematic diagram, wherein the first row indicates manual segmentation as a result, the second row is the segmentation result using distance field fusion method, third Row is the segmentation result using label fusion method, and each column indicate a test image.
Specific embodiment
The invention will be further described with the following Examples.
Embodiment 1.
A kind of MR image hippocampus dividing method based on distance field fusion, this method are based on two kinds of hypothesis:
I, MR image block and DF image block are located on two non-linearity manifolds, any one MR image block can be where it Neighbour's sample in the local space of manifold is linearly expressed；
II, under local constraint, the mapping of MR manifold to DF manifold is similar to a differomorphism mapping.
The MR image hippocampus dividing method based on distance field fusion carries out as follows:
(1) to initial MR test image normalized to be split, skull and biased field is removed, normalized is obtained MR test image afterwards.Specifically MR test image is normalized using gray scale normalization method, is gone with BET algorithm Except skull, biased field is removed with N4 algorithm.
(2) by the MR training set image and training set label corresponding with MR training set image in pre-prepd training set Image registration is to MR test image, so that MR training set image, training set label image and MR test image are on spatial position Alignment.It is specifically to use DRAMMS software tool by the MR training set image and training set labeled graph in training set in step (2) As being registrated to MR test image.
(3) range conversion is carried out to the training set label image after step (2) registration, obtains the instruction of training set label image Practice the value of training set distance field DF corresponding to collection distance field DF, point x are as follows:
Wherein C indicates the boundary of segmentation object, and B indicates the nearest point of distance x, and B ∈ C, dist (x, B) indicate point x Euclidean distance between point B.
(4) to the point x in MR test image, an image block x is taken centered on xMR, xMRIt is converted to a column vectorIt is used as the feature of point x, wherein m representation vectorDimension.
(5) a search window choosing is defined centered on point x respectively on training set MR image, training set distance field DF image Take image blockWithConstruct MR dictionary
With DF dictionary N is the size of dictionary.
(6) based on the assumption that I, with MR dictionary DMRLocal linear indicates MR test sampleDictionary DMRWeight coefficient vectorIt is solved with LAE method, expression is as follows:
Indicate test sampleIn dictionary DMRIn k neighbour.
(7) step (6) resulting dictionary D is usedMRWeight coefficient vectorLinear combination DF dictionary DDFIn sample, obtain Test sampleDF forecast image block vectorAnd handleIt is converted into image block xDF；Specifically:
Based on the assumption that I, it can obtain:
Based on the assumption that II, it can obtain:
Since f is local linear, therefore:
?It is converted into image block xDF。
(8) each point in MR test image is repeated to operate according to step (4)-(7), obtains the DF value of each point；
It is indicated centered on point x, size and x with P (x)DFEqual image block, for any point u in P (x), weight Are as follows:
The DF value of point x are as follows:
Indicate the DF value for the point x that the image block centered on point u is predicted.
(9) to the DF forecast image progress threshold process come is predicted, mark corresponding to each point of test image is obtained Number, specifically:
The distance field as defined in formula (1) is it is found that the label of point x can be exported by formula (8):
It indicates that the point belongs to segmentation object, and background is indicated marked as 0 marked as 1, hippocampus point is obtained according to segmentation object Cut image.
Preferably, above-mentioned steps (4) are specifically to extract the feature of each point in MR test image, are used as test specimens This.
Preferably, above-mentioned steps (5) are specifically that local search window is used in MR training set image and DF training set image Construct MR dictionary and DF dictionary.
Preferably, above-mentioned steps (8) are specifically to handle the DF forecast image block weighted average of overlapping, final to obtain each The DF predicted value of point.
The method of the present invention provides a kind of MR image hippocampus dividing method based on distance field fusion, and the instruction of priori is utilized Practice collection MR image, training set distance field DF image, is handled using distance field, the hippocampus in the accurate Ground Split MR image of energy Body can be used for the diagnosis of the mental diseases such as Alzheimer's disease.
Embodiment 2.
It is according to being tested in database with 35 groups of MR brain datas for including to verify the validity of the method for the present invention Card.The data of every group of MR brain data include the T1 weighted MR image and corresponding hippocampus label image of same patient, In select 20 groups of data at random as training set, remaining 15 groups of data is as test set.Test set object is only with it in experiment MR image does not use the hippocampus label image in corresponding database.
The present invention is based on the MR image hippocampus dividing method of distance field fusion, this method is based on two kinds of hypothesis:
I, MR image block and DF image block are located on two non-linearity manifolds, any one MR image block can be where it Neighbour's sample in the local space of manifold is linearly expressed；
II, under local constraint, the mapping of MR manifold to DF manifold is similar to a differomorphism mapping.
The MR image hippocampus dividing method based on distance field fusion carries out as follows:
(1) to initial MR test image normalized to be split, skull and biased field is removed, normalized is obtained MR test image afterwards.Specifically MR test image is normalized using gray scale normalization method, is gone with BET algorithm Except skull, biased field is removed with N4 algorithm.
(2) by training set MR training set image and training set label image corresponding with MR training set image be registrated to MR test image, so that MR training set image, training set label image are aligned on spatial position with MR test image.Step (2) in be specifically using DRAMMS software tool by training set MR training set image and training set label image be registrated to MR Test image.
(3) range conversion is carried out to the training set label image after step (2) registration, obtains the instruction of training set label image Practice the value of training set distance field DF corresponding to collection distance field DF, point x are as follows:
Wherein C indicates the boundary of segmentation object, and B indicates the nearest point of distance x, and B ∈ C, dist (x, B) indicate point x Euclidean distance between point B.
(4) to the point x in MR test image, an image block x is taken centered on xMR, xMRIt is converted to a column vectorIt is used as the feature of point x, wherein m representation vectorDimension.
(5) a search window choosing is defined centered on point x respectively on training set MR image, training set distance field DF image Take image blockWithConstruct MR dictionary
With DF dictionary N is the size of dictionary.
(6) based on the assumption that I, with MR dictionary DMRLocal linear indicates MR test sampleDictionary DMRWeight coefficient vectorIt is solved with LAE method, expression is as follows:
Indicate test sampleIn dictionary DMRIn k neighbour.
(7) step (6) resulting dictionary D is usedMRWeight coefficient vectorLinear combination DF dictionary DDFIn sample, obtain Test sampleDF forecast image block vectorAnd handleIt is converted into image block xDF；Specifically:
Based on the assumption that I, it can obtain:
Based on the assumption that II, it can obtain:
Since f is local linear, therefore:
?It is converted into image block xDFIt can be obtainedThe distance field image block x of predictionDF。
(8) each point in MR test image is repeated to operate according to step (4)-(7), obtains the DF value of each point；
It is indicated centered on point x, size and x with P (x)DFEqual image block, for any point u in P (x), weight Are as follows:
The DF value of point x are as follows:
Indicate the DF value for the point x that the image block centered on point u is predicted.
(9) to the DF forecast image progress threshold process come is predicted, mark corresponding to each point of test image is obtained Number, specifically:
The distance field as defined in formula (1) is it is found that the label of point x can be exported by formula (8):
It indicates that the point belongs to segmentation object, and background is indicated marked as 0 marked as 1, hippocampus point is obtained according to segmentation object Cut image.
By evaluating the accurate of segmentation result with DSC (Dice similarity coefficient, likeness coefficient) Property, consider two factors: (1) overall performance of this method；(2) compared with label fusion method (abbreviation LF), distance field has Effect property.
The experimental result of the method for the present invention hereinafter referred to as DFF method is as follows:
Fig. 4 is for the method for the present invention with different size of search window to the segmentation result box traction substation of 15 test images；Wherein, Fig. 4 (a) is the segmentation result box traction substation for right hippocampus, and Fig. 4 (b) is the segmentation result box traction substation for left hippocampus.
Fig. 5 is the segmentation result comparison diagram with label fusion method and distance field fusion method to 15 test images, In, Fig. 5 (a) is the segmentation result comparison diagram for right hippocampus, and Fig. 5 (b) is the segmentation result comparison diagram for left hippocampus. It can be seen that the DSC of the result of DFF segmentation is obviously better than LF method.
Fig. 6 is the coronal-plane of the segmentation result in four test images using distance field fusion method and label fusion method Schematic diagram, wherein the first row indicates manual segmentation as a result, the second row is the segmentation result using distance field fusion method, third Row is the segmentation result using label fusion method, and each column indicate a test image.It can be seen that the result ratio of DFF method The result precision of LF method is high.
The above result shows that can accurate Ground Split MR brain using the MR image hippocampus dividing method merged based on distance field Hippocampus in portion's image, and then the diagnosis of the mental diseases such as Alzheimer's disease can be used for.Finally it should be noted that more than Embodiment is merely illustrative of the technical solution of the present invention rather than limiting the scope of the invention, although referring to preferred embodiment The present invention is explained in detail, those skilled in the art should understand that, technical solution of the present invention can be carried out Modification or equivalent replacement, without departing from the spirit and scope of technical solution of the present invention.

Claims (5)

1. a kind of MR image hippocampus dividing method based on distance field fusion, it is characterised in that:
The MR image hippocampus dividing method based on distance field fusion is based on two kinds of hypothesis:
I, MR image block and DF image block are located on two non-linearity manifolds, the office of any one MR image block manifold where it Neighbour's sample in portion space is linearly expressed；
II, under local constraint, the mapping of MR manifold to DF manifold is similar to a differomorphism mapping；
The MR image hippocampus dividing method based on distance field fusion carries out as follows:
(1) to initial MR test image normalized to be split, skull and biased field are removed, after obtaining normalized MR test image；
(2) by the MR training set image and training set label image corresponding with MR training set image in pre-prepd training set It is registrated to MR test image, so that MR training set image, training set label image are aligned on spatial position with MR test image；
(3) range conversion is carried out to the training set label image after step (2) registration, obtains the DF training of training set label image Collect distance field, the value of DF training set distance field corresponding to point x are as follows:
Wherein C indicates the boundary of segmentation object, and B indicates the nearest point of distance x, and B ∈ C, dist (x, B) indicate point x and point B Between Euclidean distance；
(4) to the point x in MR test image, an image block x is taken centered on xMR, xMRIt is converted to a column vectorIt is used as the feature of point x, obtains test sampleWherein m representation vectorDimension；
(5) search window is defined centered on point x respectively on MR training set image, DF training set distance field image choose figure As blockWithConstruct MR dictionaryWith DF dictionary N is the size of dictionary；
(6) based on the assumption that I, with MR dictionary DMRLocal linear indicates MR test sampleDictionary DMRWeight coefficient vectorIt is solved with LAE method, expression is as follows:
Indicate test sampleIn dictionary DMRIn k neighbour；
(7) step (6) resulting dictionary D is usedMRWeight coefficient vectorLinear combination DF dictionary DDFIn sample, obtain test specimens ThisDF forecast image block vectorAnd handleIt is converted into image block xDF；Specifically:
Based on the assumption that I, obtains:
Based on the assumption that II, obtains:
Since f is local linear, therefore:
?It is converted into image block xDF；
(8) each point in MR test image is repeated to operate according to step (4)-(7), obtains the DF training set distance of each point The predicted value of field；
Specific steps are as follows:
It is indicated centered on point x, size and x with P (x)DFEqual MR image block, for any point u in P (x), weight Are as follows:
The value of the DF training set distance field of point x are as follows:
Indicate the value of the DF training set distance field for the point x that the DF image block centered on point u is predicted；
(9) to the DF forecast image progress threshold process come is predicted, label corresponding to each point of test image is obtained, is had Body is:
The distance field as defined in formula (1) knows that the label of point x is exported by formula (8):
It indicates that the point belongs to segmentation object, and background is indicated marked as 0 marked as 1, hippocampus segmentation figure is obtained according to segmentation object Picture.
2. a kind of MR image hippocampus dividing method based on distance field fusion according to claim 1, it is characterised in that: The step (1) is specifically that MR test image is normalized using gray scale normalization method, removes cranium with BET algorithm Bone removes biased field with N4 algorithm.
3. a kind of MR image hippocampus dividing method based on distance field fusion according to claim 2, it is characterised in that: The step (2) be specifically using DRAMMS software tool by training set MR training set image and training set label image match Standard arrives MR test image.
4. a kind of MR image hippocampus dividing method based on distance field fusion according to claim 3, it is characterised in that: The step (5) is specifically to construct MR word using local search window in MR training set image and the middle of DF training set distance field image Allusion quotation and DF dictionary.
5. a kind of MR image hippocampus dividing method based on distance field fusion according to claim 4, it is characterised in that: The step (8) is specifically to handle the DF forecast image block weighted average of overlapping, finally obtain the DF training set of each point away from The predicted value left the theatre.
CN201510539859.4A 2015-08-30 2015-08-30 MR image hippocampus dividing method based on distance field fusion Active CN105205810B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201510539859.4A CN105205810B (en) 2015-08-30 2015-08-30 MR image hippocampus dividing method based on distance field fusion

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201510539859.4A CN105205810B (en) 2015-08-30 2015-08-30 MR image hippocampus dividing method based on distance field fusion

Publications (2)

Publication Number Publication Date
CN105205810A CN105205810A (en) 2015-12-30
CN105205810B true CN105205810B (en) 2018-12-11

Family

ID=54953472

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201510539859.4A Active CN105205810B (en) 2015-08-30 2015-08-30 MR image hippocampus dividing method based on distance field fusion

Country Status (1)

Country Link
CN (1) CN105205810B (en)

Families Citing this family (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106530321B (en) * 2016-10-28 2019-07-12 南方医科大学 A kind of multichannel chromatogram image partition method based on direction and scale description
CN106780532A (en) * 2016-12-30 2017-05-31 陕西海泰电子有限责任公司 A kind of intermodulation frequency equation method for quick towards DF figures
CN107424145A (en) * 2017-06-08 2017-12-01 广州中国科学院软件应用技术研究所 The dividing method of nuclear magnetic resonance image based on three-dimensional full convolutional neural networks
CN107292346B (en) * 2017-07-05 2019-11-15 四川大学 A kind of MR image hippocampus partitioning algorithm based on Local Subspace study
CN107392907A (en) * 2017-09-01 2017-11-24 上海理工大学 Parahippocampal gyrus function division method based on tranquillization state FMRI
CN108937934A (en) * 2018-05-07 2018-12-07 武汉科技大学 A kind of detection of Brain MRI hippocampus and dividing method based on deep learning
CN108596924A (en) * 2018-05-17 2018-09-28 南方医科大学 A kind of MR prostate image partition methods based on distance field fusion and ellipsoid priori
CN109886972A (en) * 2019-01-24 2019-06-14 山西大学 A kind of brain magnetic resonance image partition method based on multilayer dictionary

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1952981A (en) * 2005-07-13 2007-04-25 西门子共同研究公司 Method for knowledge based image segmentation using shape models
CN104794739A (en) * 2015-05-03 2015-07-22 南方医科大学 Method for predicting CT (computerized tomography) image from MR (magnetic resonance) image on the basis of combination of corresponding partial sparse points

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8805045B2 (en) * 2010-10-12 2014-08-12 Siemens Aktiengesellschaft Interaction method for regions of-interest in time series images

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1952981A (en) * 2005-07-13 2007-04-25 西门子共同研究公司 Method for knowledge based image segmentation using shape models
CN104794739A (en) * 2015-05-03 2015-07-22 南方医科大学 Method for predicting CT (computerized tomography) image from MR (magnetic resonance) image on the basis of combination of corresponding partial sparse points

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
Adaptive Segmentation of Vertebral Bodies from Sagittal MR Images Based on Local Spatial Information and Gaussian Weighted Chi-Square Distance;Qian Zheng 等;《Society for Imaging Informatics in Medicine》;20121113;第578-593 *
Combination Strategies in Multi-Atlas Image Segmentation: Application to Brain MR Data;Xabier Artaechevarria 等;《IEEE TRANSACTIONS ON MEDICAL IMAGING》;20090831;第28卷(第8期);第1266-1277页 *
基于多权重概率图谱的脑部图像分割;张雷 等;《南方医科大学学报》;20150812;第35卷(第8期);第1143-1148页 *

Also Published As

Publication number Publication date
CN105205810A (en) 2015-12-30

Similar Documents

Publication Publication Date Title
CN105205810B (en) MR image hippocampus dividing method based on distance field fusion
Zhang et al. MI-UNet: multi-inputs UNet incorporating brain parcellation for stroke lesion segmentation from T1-weighted magnetic resonance images
ES2595210T3 (en) Method and apparatus for processing medical images
Devi et al. Neonatal brain MRI segmentation: A review
Toga et al. Maps of the brain
Li et al. A hybrid approach to automatic clustering of white matter fibers
Ceritoglu et al. Large deformation diffeomorphic metric mapping registration of reconstructed 3D histological section images and in vivo MR images
O’Donnell et al. White matter tract clustering and correspondence in populations
US20170069086A1 (en) Method and device for constructing brain templates
CN110148145B (en) Image target area extraction method fusing boundary information and application
Nowinski et al. Three-dimensional interactive and stereotactic human brain atlas of white matter tracts
CN111931811A (en) Calculation method based on super-pixel image similarity
CN104700120A (en) Method for extracting and classifying fMRI features based on adaptive entropy algorithm for projection clustering (APEC)
CN111681230A (en) System and method for scoring high-signal of white matter of brain
CN107093190A (en) A kind of Medical Image Registration Algorithm based on multichannel chromatogram tag fusion
Zhang et al. Deep white matter analysis: fast, consistent tractography segmentation across populations and dMRI acquisitions
CN110458869B (en) Registration method and system for brain image of magnetic resonance structure of newborn
Wu et al. Automatic symmetry detection from brain MRI based on a 2-channel convolutional neural network
CN103745473B (en) A kind of brain tissue extraction method
Andica et al. Automated three-dimensional major white matter bundle segmentation using diffusion magnetic resonance imaging
Bazgir et al. Kidney segmentation using 3D U-Net localized with Expectation Maximization
CN110751664B (en) Brain tissue segmentation method based on hyper-voxel matching
Amoroso et al. Automated Shape Analysis landmarks detection for medical image processing.
Huang et al. Automatic oculomotor nerve identification based on data‐driven fiber clustering
CN114782371A (en) Brain area automatic segmentation method based on artificial intelligence

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant