CN105913431A - Multi-atlas dividing method for low-resolution medical image - Google Patents
Multi-atlas dividing method for low-resolution medical image Download PDFInfo
- Publication number
- CN105913431A CN105913431A CN201610224426.4A CN201610224426A CN105913431A CN 105913431 A CN105913431 A CN 105913431A CN 201610224426 A CN201610224426 A CN 201610224426A CN 105913431 A CN105913431 A CN 105913431A
- Authority
- CN
- China
- Prior art keywords
- image
- resolution
- low
- atlas
- target image
- 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
Links
- 238000000034 method Methods 0.000 title claims abstract description 84
- 230000004927 fusion Effects 0.000 claims abstract description 10
- 238000003709 image segmentation Methods 0.000 claims description 8
- 238000005070 sampling Methods 0.000 claims description 7
- 238000012549 training Methods 0.000 claims description 5
- 241000287196 Asthenes Species 0.000 claims description 3
- 230000000644 propagated effect Effects 0.000 claims description 3
- 238000005516 engineering process Methods 0.000 abstract description 2
- 238000012545 processing Methods 0.000 abstract description 2
- 230000011218 segmentation Effects 0.000 description 44
- 210000001320 hippocampus Anatomy 0.000 description 6
- 210000004556 brain Anatomy 0.000 description 4
- 238000001228 spectrum Methods 0.000 description 4
- 241000257303 Hymenoptera Species 0.000 description 2
- 238000010191 image analysis Methods 0.000 description 2
- 238000000513 principal component analysis Methods 0.000 description 2
- 208000024827 Alzheimer disease Diseases 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000003759 clinical diagnosis Methods 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000002059 diagnostic imaging Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- PCHJSUWPFVWCPO-UHFFFAOYSA-N gold Chemical compound [Au] PCHJSUWPFVWCPO-UHFFFAOYSA-N 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000005192 partition Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 238000001356 surgical procedure Methods 0.000 description 1
- 238000003325 tomography Methods 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
-
- 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
Abstract
The invention discloses a multi-atlas dividing method for a low-resolution medical image, wherein the multi-atlas dividing method belongs to the field of image processing technology. The method comprises the steps of setting a low-resolution objective image and N high-resolution atlas images, and assuming a fact that the objective image and the atlas images are linearly registered in a same template space; and then successively performing a dividing object area cutting step; an objective image super-resolution recovering step; an image registering and label propagation step; and a label fusion step. According to the multi-atlas dividing method for the low-resolution medical image, an image super-resolution restoring method is merged into a multi-atlas dividing frame; and through improving registering precision between the high-resolution atlas images and the low-resolution to-be-divided image, dividing precision of the multi-atlas dividing method.
Description
Technical field
The present invention relates to the multichannel chromatogram dividing method of a kind of low resolution medical image, belong to technical field of image processing.
Background technology
Along with the development of medical imaging devices and universal, medical image analysis is in sides such as disease research, surgery planning, clinical diagnosises
Face plays particularly important effect.Medical image segmentation is an important step of medical image analysis, its essence is and treats point
The each pixel or the voxel that cut image carry out label, and the pixel of different attribute or voxel set different label values, thus will
Image to be split is divided into non-overlapping interest region and background area, in order to follow-up interest region is further analyzed and
Process.The generally segmentation to medical image is completed by manually marking interest region, and the advantage of the method is segmentation result accuracy
High.But, artificial dividing method takes time and effort very much, and segmentation result is the most not reproducible.In order to reach automatic and accurate
The purpose of Medical Image Segmentation, many classical image partition methods have been widely used in the segmentation of all kinds of medical image.
Dividing method based on collection of illustrative plates, the collection of illustrative plates owing to being used not only contains the half-tone information of image, and provides image
The prior informations such as the shape of middle zones of different, therefore can obtain more accurate segmentation result compared to other dividing method.Root
According to the quantity of use collection of illustrative plates, dividing method based on collection of illustrative plates can be divided into free hand drawing spectrum dividing method and multichannel chromatogram dividing method.Free hand drawing
Spectrum dividing method is owing to only with single collection of illustrative plates, just having obtained mesh after collection of illustrative plates label image being traveled to target image by registration
The segmentation result of logo image.And multichannel chromatogram dividing method uses multiple collection of illustrative plates, by registration, each collection of illustrative plates label image is traveled to
Having obtained multiple segmentations of target image after target image, these segmentation results are combined by tag fusion and obtain final dividing
Cut result.Therefore, multichannel chromatogram dividing method can be largely classified into image registration and two steps of tag fusion.With free hand drawing spectrum point
Segmentation method is compared, and multichannel chromatogram dividing method, owing to employing more collection of illustrative plates, therefore contains the individual difference letter in region to be split
Breath, and can also effectively eliminate, by tag fusion, the error that in free hand drawing spectral method, gray level image registration introduces.
Multichannel chromatogram dividing method obtains in the segmentation field of medical image at present and applies widely.But according to the report of current document
Lead, it has been found that: atlas image that (1) multichannel chromatogram dividing method uses and target image often from same image set,
I.e. there is identical picture quality (resolution ratio).(2) multichannel chromatogram dividing method is on different data sets, to identical brain structure
Splitting, the segmentation result obtained differs huge in accuracy.Such as the segmentation of hippocampus, current main flow
Multichannel chromatogram method can obtain segmentation precision at high-resolution MRI data set (such as ADNI data set) and (use Dice
Value metric, Dice value the biggest expression precision is the highest) it is 0.9;And at low resolution image data collection (such as mTLE data set),
The segmentation precision obtained is only 0.72.
In the actual application of multichannel chromatogram dividing method, owing to collection of illustrative plates constructs in advance, full resolution pricture is generally used to be configured to
To high-quality collection of illustrative plates;But being affected by the factor such as image acquisition time, collecting device, the quality of target image cannot ensure,
The target image of low resolution can be obtained in the case of a lot.Therefore, how to split low resolution target image with high-resolution collection of illustrative plates,
The actual application of multichannel chromatogram dividing method has important researching value.
In view of this, this is studied by the present inventor, develops the multichannel chromatogram segmentation side of a kind of low resolution medical image specially
Method, this case thus produces.
Summary of the invention
It is an object of the invention to provide the multichannel chromatogram dividing method of a kind of low resolution medical image, in conjunction with superresolution restoration method
Improve dividing method based on multichannel chromatogram, propose accurate, the low resolution medical science brain image segmentation method of robust.
To achieve these goals, the solution of the present invention is:
The multichannel chromatogram dividing method of low resolution medical image, comprises the steps:
Step one, given low resolution target image Id, N number of high-resolution atlas image Ai=(Ii,Li), i=1,2 ..., N, its
Middle IiRepresent i-th gray level image, LiRepresent the label image that i-th gray level image is corresponding, and suppose target image IdAnd
Atlas image Ai=(Ii,Li), i=1,2 ..., N, the most linearly it is registrated to same templatespace;
Step 2, the cutting of cutting object region: scan all atlas image Ai=(Ii,Li), i=1,2 ..., N, in label image,
(x, y, z) position, if extending along on three coordinate directions to find out the minimum and maximum three-dimensional coordinate of cutting object in each collection of illustrative plates
Dry voxel is so that the cutting object in target image can be included;
Step 3, target image superresolution restoration: use image super-resolution restoration methods based on rarefaction representation to target figure
As IdCarry out superresolution restoration, by low resolution target image IdRevert to the higher resolution image I of identical contenth;
Step 4, image registration and label are propagated: to each atlas image Ai=(Ii,Li), i=1,2 ..., N., independently it
And superresolution restoration after target image IhRegistrate, by registration, obtain each atlas image
Ai=(Ii,Li), i=1,2 ..., N. to target image IdDeformation Field, utilize above-mentioned Deformation Field by atlas image
Ai=(Ii,Li), i=1,2 ..., the label image of N. travels to target image space, then obtains each atlas image to target image
Segmentation result;
Step 5, tag fusion: use most of voting method to carry out tag fusion, for each voxel of target image,
Its label value is determined by that label value that each atlas image correspondence position occurrence number is most.
As preferably: image super-resolution restoration methods described in step 3 particularly as follows:
Utilize one group of high-definition picture { Ii, i=1,2 ..., N} as training set by low resolution target image IdRevert to identical interior
The higher resolution image I heldh;
High-definition picture IhTo low-resolution image IdThe process that degrades typically represent by following Mathematical Modeling
Id=LMBIh, (1)
Here B and LMRepresent respectively image IhCarry out fuzzy and down-sampling;
Image super-resolution method constructs high resolution graphics image set { I first with the model that degrades (1)i, i=1,2 ..., corresponding low of N}
Resolution chart image set, is designated asThen, high-definition picture block dictionary D is constructed according to the two image seth
And the low-resolution image block dictionary D of correspondenced, for low-resolution image IdEach image block pd, utilize sparse table
Representation model
Low-resolution image block p can be obtaineddIn low-resolution dictionary DdIn expression: pd=Ddα, it is assumed that corresponding high-resolution
Image block phAt high-definition picture block dictionary DhIn there is identical representation (i.e. having identical expression factor alpha), then
High-definition picture block p can be calculatedh=Dhα, by low-resolution image IdAll image blocks recover correspondence high-resolution
After rate image block, initial full resolution pricture I can be obtained through assembly0, by an overall regularization model
Obtain final full resolution pricture Ih。
As preferably, the registration described in step 4 is to be obtained by the nonlinear algorithm driven based on crosscorrelation, available public generation
Code kit ANTs realizes.
The multichannel chromatogram dividing method of low resolution medical image of the present invention, is dissolved into many figures by Image Super-resolution restoration methods
Among spectrum segmentation framework, by improving the registration accuracy between high-resolution atlas image and low resolution image to be split, thus carry
The segmentation precision of high multichannel chromatogram dividing method.Improve multichannel chromatogram dividing method the highest to low resolution target image segmentation accuracy
Situation.
Below in conjunction with drawings and the specific embodiments, the present invention is described in further detail.
Accompanying drawing explanation
Fig. 1 is the multichannel chromatogram dividing method flow chart of the low resolution medical science brain image of the present embodiment;
Fig. 2 is one of them tested MRI of the present embodiment, hippocampus segmentation image and the low-resolution image of structure
(first and second images are a tested magnetic resonance gray level image from ADNI database and the segmentation of corresponding hippocampus
Figure, the 3rd image is the low-resolution image constructed by obfuscation and down-sampling);
Fig. 3 is the segmentation result example of the present embodiment, and first row is former 3.0T image and corresponding artificial segmentation (as goldstandard);
Its segmentation is obtained by secondary series is obfuscation and down-sampling obtains low resolution target image and original multichannel chromatogram dividing method
Segmentation result;3rd row be superresolution restoration after target image and the present embodiment the segmentation result that obtains of method is proposed.
Detailed description of the invention
The multichannel chromatogram dividing method of low resolution medical image, comprises the steps:
Step one, given low resolution target image Id, N number of high-resolution atlas image Ai=(Ii,Li), i=1,2 ..., N, its
Middle IiRepresent i-th gray level image, LiRepresent the label image that i-th gray level image is corresponding, and suppose target image IdAnd
Atlas image Ai=(Ii,Li), i=1,2 ..., N, the most linearly it is registrated to same templatespace;
Step 2, the cutting of cutting object region:
As it is shown in figure 1, for reducing amount of calculation, first the region comprising cutting object is cut down.Due to target
Image IdAnd atlas image Ai=(Ii,Li), i=1,2 ..., N, the most linearly it is registrated to a templatespace, has therefore been split
Object position in each image is the most similar, scans all atlas image Ai=(Ii,Li), i=1,2 ..., N, in label image (i.e.
Scan all segmentation results), (x, y, z) position, along three to find out the minimum and maximum three-dimensional coordinate of cutting object in each collection of illustrative plates
Several voxels are extended so that the cutting object in target image can be included on individual coordinate direction.For description side
Just, target image and atlas image after cutting are still designated as IdAnd Ai=(Ii,Li), i=1,2 ..., N.
Step 3, target image superresolution restoration:
Use image super-resolution restoration methods based on rarefaction representation to target image IdCarry out superresolution restoration: utilize one group
High-definition picture { Ii, i=1,2 ..., N} as training set by low resolution target image IdRevert to the more high-resolution of identical content
Rate image Ih。
High-definition picture IhTo low-resolution image IdThe process that degrades typically represent by following Mathematical Modeling
Id=LMBIh, (1)
Here B and LMRepresent respectively image IhCarry out fuzzy and down-sampling.
Image super-resolution restoration methods constructs high resolution graphics image set { I first with the model that degrades (1)i, i=1,2 ..., N} is corresponding
Low-resolution image collection, be designated asThen, high-definition picture block is constructed according to the two image set
Dictionary DhAnd the low-resolution image block dictionary D of correspondenced.For low-resolution image IdEach image block pd, utilize
Sparse representation model
Low-resolution image block p can be obtaineddIn low-resolution dictionary DdIn expression: pd=Ddα.Assuming that corresponding high score
Resolution image block phAt high-definition picture block dictionary DhIn there is identical representation (i.e. having identical expression factor alpha),
Then high-definition picture block p can be calculatedh=Dhα.By low-resolution image IdAll image blocks recover correspondence height
After image in different resolution block, initial full resolution pricture I can be obtained through assembly0, by an overall regularization model
Obtain final full resolution pricture Ih。
Step 4, image registration and label are propagated:
To each atlas image Ai=(Ii,Li), i=1,2 ..., N., independently the target figure after they and superresolution restoration
As IhRegistrating, registration is to be obtained by the nonlinear algorithm driven based on crosscorrelation, available common code kit ANTs
Realize.By registration, obtain each atlas image Ai=(Ii,Li), i=1,2 ..., N. to target image IdDeformation Field.Profit
By above-mentioned Deformation Field by atlas image Ai=(Ii,Li), i=1,2 ..., the label image of N. travels to target image space, then obtains
Each atlas image segmentation result to target image.
Step 5, tag fusion:
Using most of voting method to carry out tag fusion, for each voxel of target image, its label value is by each collection of illustrative plates
That label value decision that image correspondence position occurrence number is most.
The multichannel chromatogram dividing method of the low resolution medical image described in the present embodiment, is dissolved into many by Image Super-resolution restoration methods
Among collection of illustrative plates segmentation framework, by improving the registration accuracy between high-resolution atlas image and low resolution image to be split, thus
Improve the segmentation precision of multichannel chromatogram dividing method.Improve multichannel chromatogram dividing method to low resolution target image segmentation accuracy not
High situation.
In the present embodiment, use from the 30 of Alzheimer disease image center database (ADNI, adni.loni.ucla.edu)
Individual tested 3.0T MRI verifies the multichannel chromatogram dividing method of above-mentioned low resolution medical image, and cutting object is brain
Hippocampus.For each tested, the method for manual segmentation is first used to obtain the hippocampus segmentation of correspondence;By image blurringization and
Down-sampling, the low-resolution image that structure is corresponding.Fig. 2 shows one of them tested MRI, hippocampus segmentation figure
Picture and the low-resolution image of structure, for convenience of showing, have chosen one of them two dimension tomography and be indicated.In these data
In, the low resolution image of structure, is adopted as target image to be split, original 3.0T MRI as training data
The method verifying proposition by the mode staying a checking.
Parameter sets:
Dictionary size in image super-resolution restoration methods, random in each training data chooses 100 image blocks, word
Allusion quotation size is 100*29 that is 2900.The dictionary structure of low-resolution image, when extracting feature with principal component analysis (PCA),
We set the ratio of keeping characteristics dimension as 10%, and the size of the image block extracted in low resolution is set as 3 × 3 × 3.High score
The scale factor of resolution image and low-resolution image is 2 × 2 × 2 for ADNI data setting.
Segmentation result:
Calculate segmentation precision by Dice value, be defined as follows, given artificial segmentation result E (as goldstandard) and from
Dynamic segmentation result F,
Here V (X) represents the volume of segmentation result X.
Multichannel chromatogram dividing method based on superresolution restoration described in the present embodiment and original multichannel chromatogram dividing method are compared
Relatively.
The segmentation that multichannel chromatogram dividing method described in the segmentation result of the multichannel chromatogram dividing method that table 1. is original and the present embodiment obtains
The Dice mean value of result and standard deviation.
Table 1 lists Dice mean value and the standard deviation of segmentation result, and compares original multichannel chromatogram with the inspection of double sample t
The otherness of the segmentation result that the multichannel chromatogram dividing method described in the segmentation result of dividing method and the present embodiment obtains.Experimental result
Display, in level of significance α value 10-5In the case of, the multichannel chromatogram dividing method described in the present embodiment to be significantly better than original
Multichannel chromatogram dividing method.
Fig. 3 illustrates a tested segmentation situation at random, and first row is former 3.0T image and corresponding manually splitting (as gold
Standard);It is split by secondary series is obfuscation and down-sampling obtains low resolution target image and original multichannel chromatogram dividing method
The segmentation result obtained;3rd row be superresolution restoration after target image and the present embodiment propose method obtain segmentation knot
Really.It can be seen that: (1) uses superresolution restoration method to improve quality (2,3 two of contrast the first row of image
Figure);(2) combine superresolution restoration method and segmentation result that multichannel chromatogram dividing method obtains is more accurately (by the second row 2,3
Two figures contrast 1 figure manually split can be seen that).
Conclusion: by combining superresolution restoration method, low resolution target image is carried out superresolution restoration, can actually carry
The accuracy of high multichannel chromatogram dividing method and robustness.
Above-described embodiment and the product form of the graphic and non-limiting present invention and style, the ordinary skill people of any art
Member is suitably changed what it did or modifies, and all should be regarded as the patent category without departing from the present invention.
Claims (3)
1. the multichannel chromatogram dividing method of low resolution medical image, it is characterised in that comprise the steps:
Step one, given low resolution target image Id, N number of high-resolution atlas image Ai=(Ii,Li), i=1,2 ..., N, wherein IiTable
Show i-th gray level image, LiRepresent the label image that i-th gray level image is corresponding, and suppose target image IdAnd figure spectrogram
As Ai=(Ii,Li), i=1,2 ..., N, the most linearly it is registrated to same templatespace;
Step 2, the cutting of cutting object region: scan all atlas image Ai=(Ii,Li), i=1,2 ..., N, in label image, find out
In each collection of illustrative plates, (x, y, z) position extend several on three coordinate directions to the minimum and maximum three-dimensional coordinate of cutting object
Voxel is so that the cutting object in target image can be included;
Step 3, target image superresolution restoration: use image super-resolution restoration methods based on rarefaction representation to target image IdEnter
Row superresolution restoration, by low resolution target image IdRevert to the higher resolution image I of identical contenth;
Step 4, image registration and label are propagated: to each atlas image Ai=(Ii,Li), i=1,2 ..., N., independently they and
Target image I after superresolution restorationhRegistrate, by registration, obtain each atlas image
Ai=(Ii,Li), i=1,2 ..., N. to target image IdDeformation Field, utilize above-mentioned Deformation Field by atlas image
Ai=(Ii,Li), i=1,2 ..., the label image of N. travels to target image space, then obtains each atlas image to target image
Segmentation result;
Step 5, tag fusion: use most of voting method to carry out tag fusion, for each voxel of target image, its
Label value is determined by that label value that each atlas image correspondence position occurrence number is most.
2. the multichannel chromatogram dividing method of low resolution medical image as claimed in claim 1, it is characterised in that: described in step 3
Image super-resolution restoration methods particularly as follows:
Utilize one group of high-definition picture { Ii, i=1,2 ..., N} as training set by low resolution target image IdRevert to identical content
Higher resolution image Ih;
High-definition picture IhTo low-resolution image IdThe process that degrades typically represent by following Mathematical Modeling
Id=LMBIh, (1)
Here B and LMRepresent respectively image IhCarry out fuzzy and down-sampling;
Image super-resolution restoration methods constructs high resolution graphics image set { I first with the model that degrades (1)i, i=1,2 ..., corresponding low of N}
Resolution chart image set, is designated asThen, high-definition picture block dictionary D is constructed according to the two image seth
And the low-resolution image block dictionary D of correspondenced, for low-resolution image IdEach image block pd, utilize sparse table
Representation model
Low-resolution image block p can be obtaineddIn low-resolution dictionary DdIn expression: pd=Ddα, it is assumed that corresponding high-resolution
Image block phAt high-definition picture block dictionary DhIn there is identical representation, then can calculate high-definition picture block
ph=Dhα, by low-resolution image IdAll image blocks recover correspondence high-definition picture block after, can through assembly
To obtain initial full resolution pricture I0, by an overall regularization model
Obtain final full resolution pricture Ih。
3. the multichannel chromatogram dividing method of low resolution medical image as claimed in claim 1, it is characterised in that: described in step 4
Registration is to be obtained by the nonlinear algorithm driven based on crosscorrelation.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610224426.4A CN105913431A (en) | 2016-04-12 | 2016-04-12 | Multi-atlas dividing method for low-resolution medical image |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610224426.4A CN105913431A (en) | 2016-04-12 | 2016-04-12 | Multi-atlas dividing method for low-resolution medical image |
Publications (1)
Publication Number | Publication Date |
---|---|
CN105913431A true CN105913431A (en) | 2016-08-31 |
Family
ID=56745907
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610224426.4A Pending CN105913431A (en) | 2016-04-12 | 2016-04-12 | Multi-atlas dividing method for low-resolution medical image |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105913431A (en) |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107093190A (en) * | 2017-04-17 | 2017-08-25 | 哈尔滨理工大学 | A kind of Medical Image Registration Algorithm based on multichannel chromatogram tag fusion |
CN107766857A (en) * | 2017-10-17 | 2018-03-06 | 天津大学 | The vision significance detection algorithm propagated based on graph model structure with label |
CN108564607A (en) * | 2018-04-08 | 2018-09-21 | 华中科技大学苏州脑空间信息研究院 | Three-dimensional brain map data set space calibration method and system |
CN109242865A (en) * | 2018-09-26 | 2019-01-18 | 上海联影智能医疗科技有限公司 | Medical image auto-partition system, method, apparatus and storage medium based on multichannel chromatogram |
CN109523512A (en) * | 2018-10-17 | 2019-03-26 | 哈尔滨理工大学 | A kind of Automatic medical image segmentation method based on multichannel chromatogram tag fusion |
CN110853043A (en) * | 2019-11-21 | 2020-02-28 | 北京推想科技有限公司 | Image segmentation method and device, readable storage medium and electronic equipment |
CN110889816A (en) * | 2019-11-07 | 2020-03-17 | 北京量健智能科技有限公司 | Image segmentation method and device |
CN111915622A (en) * | 2020-07-09 | 2020-11-10 | 沈阳先进医疗设备技术孵化中心有限公司 | Training method and device of image segmentation network model and image segmentation method and device |
CN113935928A (en) * | 2020-07-13 | 2022-01-14 | 四川大学 | Rock core image super-resolution reconstruction based on Raw format |
US11227390B2 (en) | 2018-09-26 | 2022-01-18 | Shanghai United Imaging Intelligence Co., Ltd. | Systems and methods for image processing |
CN116228786A (en) * | 2023-05-10 | 2023-06-06 | 青岛市中心医院 | Prostate MRI image enhancement segmentation method, device, electronic equipment and storage medium |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102646268A (en) * | 2012-02-22 | 2012-08-22 | 中国科学院自动化研究所 | Magnetic resonance image brain structure automatic dividing method based on statistics multi-map registration optimization |
CN103098090A (en) * | 2011-12-21 | 2013-05-08 | 中国科学院自动化研究所 | Multiparameter three-dimensional magnetic resonance imaging brain tumor partition method |
US20140226889A1 (en) * | 2013-02-11 | 2014-08-14 | General Electric Company | Systems and methods for image segmentation using target image intensity |
US20150086096A1 (en) * | 2013-09-20 | 2015-03-26 | General Electric Company | Systems and methods for image segmentation using a deformable atlas |
-
2016
- 2016-04-12 CN CN201610224426.4A patent/CN105913431A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103098090A (en) * | 2011-12-21 | 2013-05-08 | 中国科学院自动化研究所 | Multiparameter three-dimensional magnetic resonance imaging brain tumor partition method |
CN102646268A (en) * | 2012-02-22 | 2012-08-22 | 中国科学院自动化研究所 | Magnetic resonance image brain structure automatic dividing method based on statistics multi-map registration optimization |
US20140226889A1 (en) * | 2013-02-11 | 2014-08-14 | General Electric Company | Systems and methods for image segmentation using target image intensity |
US20150086096A1 (en) * | 2013-09-20 | 2015-03-26 | General Electric Company | Systems and methods for image segmentation using a deformable atlas |
Non-Patent Citations (4)
Title |
---|
ANDREA RUEDA 等: "Single-image super-resolution of brain MR images using overcomplete dictionaries", 《MEDICAL IMAGE ANALYSIS》 * |
周建军 等: "《海战场侦察技术概论》", 31 January 2013 * |
潘红: "脑部MRI海马体三维分割算法研究", 《中国优秀硕士学位论文全文数据库》 * |
高新瑞: "《Java 3D与计算机三维动态图形网络编程设计》", 31 July 2014 * |
Cited By (20)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107093190A (en) * | 2017-04-17 | 2017-08-25 | 哈尔滨理工大学 | A kind of Medical Image Registration Algorithm based on multichannel chromatogram tag fusion |
CN107766857A (en) * | 2017-10-17 | 2018-03-06 | 天津大学 | The vision significance detection algorithm propagated based on graph model structure with label |
CN107766857B (en) * | 2017-10-17 | 2021-08-03 | 天津大学 | Visual saliency detection algorithm based on graph model construction and label propagation |
CN108564607B (en) * | 2018-04-08 | 2021-08-03 | 华中科技大学苏州脑空间信息研究院 | Three-dimensional brain atlas data set space calibration method and system |
CN108564607A (en) * | 2018-04-08 | 2018-09-21 | 华中科技大学苏州脑空间信息研究院 | Three-dimensional brain map data set space calibration method and system |
US11615535B2 (en) | 2018-09-26 | 2023-03-28 | Shanghai United Imaging Intelligence Co., Ltd. | Systems and methods for image processing |
CN109242865A (en) * | 2018-09-26 | 2019-01-18 | 上海联影智能医疗科技有限公司 | Medical image auto-partition system, method, apparatus and storage medium based on multichannel chromatogram |
US11227390B2 (en) | 2018-09-26 | 2022-01-18 | Shanghai United Imaging Intelligence Co., Ltd. | Systems and methods for image processing |
CN109523512B (en) * | 2018-10-17 | 2019-10-29 | 哈尔滨理工大学 | A kind of Automatic medical image segmentation method based on multichannel chromatogram tag fusion |
CN109523512A (en) * | 2018-10-17 | 2019-03-26 | 哈尔滨理工大学 | A kind of Automatic medical image segmentation method based on multichannel chromatogram tag fusion |
CN110889816B (en) * | 2019-11-07 | 2022-12-16 | 拜耳股份有限公司 | Image segmentation method and device |
CN110889816A (en) * | 2019-11-07 | 2020-03-17 | 北京量健智能科技有限公司 | Image segmentation method and device |
CN110853043B (en) * | 2019-11-21 | 2020-09-29 | 北京推想科技有限公司 | Image segmentation method and device, readable storage medium and electronic equipment |
CN110853043A (en) * | 2019-11-21 | 2020-02-28 | 北京推想科技有限公司 | Image segmentation method and device, readable storage medium and electronic equipment |
CN111915622A (en) * | 2020-07-09 | 2020-11-10 | 沈阳先进医疗设备技术孵化中心有限公司 | Training method and device of image segmentation network model and image segmentation method and device |
CN111915622B (en) * | 2020-07-09 | 2024-01-23 | 沈阳先进医疗设备技术孵化中心有限公司 | Training of image segmentation network model and image segmentation method and device |
CN113935928A (en) * | 2020-07-13 | 2022-01-14 | 四川大学 | Rock core image super-resolution reconstruction based on Raw format |
CN113935928B (en) * | 2020-07-13 | 2023-04-11 | 四川大学 | Rock core image super-resolution reconstruction based on Raw format |
CN116228786A (en) * | 2023-05-10 | 2023-06-06 | 青岛市中心医院 | Prostate MRI image enhancement segmentation method, device, electronic equipment and storage medium |
CN116228786B (en) * | 2023-05-10 | 2023-08-08 | 青岛市中心医院 | Prostate MRI image enhancement segmentation method, device, electronic equipment and storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN105913431A (en) | Multi-atlas dividing method for low-resolution medical image | |
CN112017198B (en) | Right ventricle segmentation method and device based on self-attention mechanism multi-scale features | |
EP3869387A1 (en) | Method and device for three-dimensional image semantic segmentation, terminal and storage medium | |
CN109410219A (en) | A kind of image partition method, device and computer readable storage medium based on pyramid fusion study | |
CN102354397B (en) | Method for reconstructing human facial image super-resolution based on similarity of facial characteristic organs | |
CN104484886B (en) | A kind of dividing method and device of MR images | |
CN103793904B (en) | Image registration device and the method for image registration thereof | |
CN109035172B (en) | Non-local mean ultrasonic image denoising method based on deep learning | |
CN109146899A (en) | CT image jeopardizes organ segmentation method and device | |
CN110197492A (en) | A kind of cardiac MRI left ventricle dividing method and system | |
US20220335600A1 (en) | Method, device, and storage medium for lesion segmentation and recist diameter prediction via click-driven attention and dual-path connection | |
US20210225000A1 (en) | Method and device for stratified image segmentation | |
US11430123B2 (en) | Sampling latent variables to generate multiple segmentations of an image | |
US11514621B2 (en) | Low-dose image reconstruction method and system based on prior anatomical structure difference | |
CN113506308A (en) | Deep learning-based vertebra positioning and spine segmentation method in medical image | |
CN112036298A (en) | Cell detection method based on double-segment block convolutional neural network | |
CN116468732A (en) | Lung CT image segmentation method and imaging method based on deep learning | |
CN103745473B (en) | A kind of brain tissue extraction method | |
CN113744275B (en) | Feature transformation-based three-dimensional CBCT tooth image segmentation method | |
Fan et al. | TR-Gan: multi-session future MRI prediction with temporal recurrent generative adversarial Network | |
DE112011104887T5 (en) | Representation of places in the medical diagnosis | |
Teixeira et al. | Adversarial data augmentation on breast MRI Segmentation | |
CN113409447B (en) | Coronary artery segmentation method and device based on multi-slice combination | |
CN113456031A (en) | Training device and prediction device of brain state prediction model and electronic equipment | |
CN114283406A (en) | Cell image recognition method, device, equipment, medium and computer program product |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into 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: 20160831 |