CN107016683A - The level set hippocampus image partition method initialized based on region growing - Google Patents
The level set hippocampus image partition method initialized based on region growing Download PDFInfo
- Publication number
- CN107016683A CN107016683A CN201710225569.1A CN201710225569A CN107016683A CN 107016683 A CN107016683 A CN 107016683A CN 201710225569 A CN201710225569 A CN 201710225569A CN 107016683 A CN107016683 A CN 107016683A
- Authority
- CN
- China
- Prior art keywords
- hippocampus
- region
- image
- level set
- pixel
- 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
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
- G06T2207/20036—Morphological image processing
-
- 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
- G06T2207/20112—Image segmentation details
- G06T2207/20156—Automatic seed setting
-
- 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
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention provides a kind of level set hippocampus image partition method initialized based on region growing, belong to field of medical image processing.This level set hippocampus image partition method initialized based on region growing, is comprised the following steps：Step S1：The clip image for including whole hippocampal formation regions is partitioned into by seed point, the gray average in seed point region is calculated；Step S2：Seed point and the similitude of seed point neighborhood territory pixel are calculated using criterion is grown, and the pixel recycling growth criterion newly obtained is calculated, until not new pixel is received；Step S3：With morphological operation, orderly hippocampus contour line is obtained.Step S4：Global gaussian probability information is incorporated, the segmentation of hippocampus is realized using improved level set algorithm, driving contour curve stops to gtoal setting and in hippocampus boundary.The present invention, which has, to improve the advantage of the precision of dividing method while segmentation result accuracy is ensured.
Description
Technical field
The invention belongs to field of medical image processing, it is related to a kind of level set hippocampus image initialized based on region growing
Dividing method.
Background technology
People's cerebral hippocampal (Hippocampus) is located at brain temporal lobe edge, belongs to cortex gyrus tissue, arched protuberance,
It is the important component of human brain, plays memory and sterically defined function.Clinical research shows, hippocampus volume morphing and
The nervous system diseases such as anomalous variation in volume and senile dementia, temporal epilepsy, schizophrenia, mild cognitive impairment have
Closely contact.Therefore, to the accurate measurement of hippocampus volume, either the functional study of human brain is still suffered to diagnosis
The state of an illness of person all has great Research Significance.
MRI, with higher contrast and resolution ratio, is widely used in human body brain as a kind of noninvasive detection methods
Portion's anatomy and pathological research, thus, the nuclear magnetic resonance (Magnetic Resonance, MR) that NMR is scanned
Image, is the Main Means for studying brain tissue anatomical structure and hippocampus function.The dividing method of hippocampus structure can in MR images
To be divided into manual segmentation, the automatic Interactive Segmentation split and fallen between.Manual segmentation is that doctor directly takes turns hippocampus
Exterior feature is sketched out in MR image slices to be come, and it needs operating personnel to carry out substantial amounts of exercise.In all dividing methods, manually
The precision of split plot design is highest, therefore often as the optimality criterion for evaluating hippocampus dividing method.But, unfortunately should
Method is relatively time-consuming, segmentation is inefficient, labor intensity is larger, and to the structure of knowledge and practical experience of trainee itself
It is required that it is higher, easily disturbed by personal subjective judgement.In actual applications, due to by equipment, skew field-effect, noise, human body
To there is contrast not high and similar to its perienchyma for hippocampal formation in the influence of the factors such as respiratory movement, Brain MR Image
The problems such as, all the Accurate Segmentation to hippocampus brings great difficulty.
In recent years, the movable contour model based on partial differential equation relies on its free flexible structure of topological sum, obtains
The favor of numerous researchers.The gradient information of multi-scale edge is added to the external energy of level set function by some research
Xiang Zhong, it is proposed that a kind of multi-scale level set dividing method based on edge information fusion, so that in the speed of service, segmentation precision
Great improvement is obtained for in terms of noise.In addition, also some researchers propose that a kind of T1 based on multichannel chromatogram registration adds
Weigh brain hippocampus Image Automatic Segmentation algorithm.By half-tone information and space bit confidence that pixel is added in registration technique
Breath so that the priori of medical image resources more effectively can be dissolved into cutting procedure by model, improve the robustness of algorithm
And registration accuracy.But, it is theoretic because the location of hippocampus is special | ▽ (Gσ* I) | → ∞ can not reach, i.e. g values
Can not be approximately 0, so that evolution curve directed overshoot border, causes over-segmentation or less divided.Therefore, only consider
The hippocampus contours segmentation method of gradient information has certain limitation.
Therefore, the present invention obtains the general profile of hippocampus using adaptive region growth algorithm and morphological operation, so
Afterwards using this contour curve as priori, the energy of curve construction evolution is carried out with reference to the gradient information and global area information of image
Equation is measured, a kind of level set hippocampus image partition method initialized based on region growing is proposed.
The content of the invention
The purpose of the present invention is there is above mentioned problem for existing technology, it is proposed that one kind is based on region growing initialization
Level set hippocampus image partition method, with can ensure segmentation result accuracy while, improve dividing method essence
The characteristics of spending.
The purpose of the present invention can be realized by following technical proposal：
A kind of level set hippocampus image partition method initialized based on region growing, it is characterised in that including following step
Suddenly：
Step S1：Original image is inputted, a seed point is chosen from the original image of whole hippocampal formations, by seed point point
The clip image for including whole hippocampal formation regions is cut out, and the gray average in seed point region is calculated, in human brain
In MR images, the area shared by hippocampus is very small, if splitting to entire image, not only increases amount of calculation, Er Qiehui
Produce the segmentation of many mistakes.Therefore, using the seed point artificially chosen as center, region of the interception comprising hippocampal formation is made
For process object, clip image need to comprising hippocampus Zone Full and remove unnecessary non-target tissues as much as possible.At this
In, we set the size of crop window as 41*41 pixels, and follow-up all processing are carried out on the clip image of setting,
Finally only need to the segmentation result on clip image to be mapped on original image, it is possible to obtain the actual profile of hippocampus；
Step S2：Seed point and the similitude of the neighborhood territory pixel of seed point 3 × 3 are calculated using criterion is grown, with identical
Or the pixel of similar gray-value is referred to the region where sub-pixel, and adaptive region is reused to the pixel newly obtained
Grow criterion to calculate, until not new pixel is received, algorithm terminates；
Step S3：The interference that the bridge joint of respective pixel point and branch are present is removed with morphological operation, obtains orderly
Hippocampus contour line.
Step S4：On the basis of obtaining orderly hippocampus contour line, to incorporate global gaussian probability information, using improved
Level set algorithm realizes the segmentation of hippocampus, and driving contour curve stops to gtoal setting and in hippocampus boundary.
It is in step sl, described in the above-mentioned level set hippocampus image partition method initialized based on region growing
The calculation formula of gray average is in seed point region：
In formula,To calculate obtained seed region gray average；(x, y) is the seed point coordinates manually chosen；I
(i, j) is the gray value at pixel (i, j) place；R is seed region.
Generally, the growth criterion of design section growth algorithm can be directly carried out with the gray average of sub-pixel.
But, as n=1 (n represents the number of pixel in seed region), hippocampus is because the presence of noise spot, the knot of growth
Just very likely there is certain deviation in fruit.At this time, it may be necessary to the gray average in sub-pixel neighborhood be obtained, with it come substitute species
Gray average in subregion, be thus avoided as much as seed point falsely drops the shadow produced with noise spot to growth result
Ring.
In the above-mentioned level set hippocampus image partition method initialized based on region growing, in step sl, from cutting
The standard deviation of pixel grey scale in image-region is cut as the threshold values of growth, the gray average in clip image regionAnd variance
σGlobalCalculation formula be：
In formula, M × N is the size of clip image.
Just can adaptively it be adjusted according to the gray feature of image in itself using such method growth threshold values, can
The blindness of artificial selection threshold values is effectively prevented from, the adaptivity of region growing can be strengthened.
In the above-mentioned level set hippocampus image partition method initialized based on region growing, the meter of described growth criterion
Calculating formula is：
In formula, ξ is standard deviation sigmaGlobalWeight.
In the above-mentioned level set hippocampus image partition method initialized based on region growing, step S3 includes following sub-step
Suddenly：
1. hole region in region growing result is filled by S3；
2. S3 the target in image is marked；
3. S3 implements opening operation to marked region；
4. S3 extracts the connected region where image center；
5. S3 obtains the profile of connected region；
6. S3 extracts the convex polygon profile of connected region.
In the above-mentioned level set hippocampus image partition method initialized based on region growing, step S4 includes following sub-step
Suddenly：
1. S4 sets up level set movements equation：
In formula, φ is level set function, and I (x) is initial pictures, and λ, ν, μ, τ are respectively length, area, canonical
The weight coefficient of item and global Gauss Distribution Fitting energy term, div is bent curvature of a curve, δεAnd H (g)ε(g) it is respectively Dirac letters
δ (z) and Heaviside function H (z) approximate expression are counted, its calculation formula is：
G is edge indicator function, and its calculation formula is：
In formula,To represent image gradient, GσThe Gaussian convolution function for being σ for standard deviation.
e1、e2For global Gauss Distribution Fitting energy term, its calculation formula is：
In formula, uiFor the global average of gray level image, σiPoor for the global criteria of gray level image, its calculation formula is：
2. S4 initializes level set function φ (x, y, t)=0 using the result of criterion algorithm and morphological operation is grown；
S4 3. initiation parameter μ, λ, ν, τ, ε, σ, Δ t；
4. S4 updates e1、e2；
5. S4 utilizes finite difference calculus, updates level set function φ；
6. whether determined level collection evolution curve meets convergence criterion to S4, if do not have, and go to step S4 4. continue calculating,
Until meeting end condition.
Compared with prior art, this is had following excellent based on the level set hippocampus image partition method that region growing is initialized
Point：The general profile of hippocampus is obtained using growth algorithm and morphological operation, then using this contour curve as priori,
Carry out the energy equation of curve construction evolution with reference to the gradient information and global area information of image, ensure segmentation result accuracy
While further improve the precision of dividing method.
Brief description of the drawings
Fig. 1 is the FB(flow block) of method involved in the present invention；
Fig. 2 is the flow chart that adaptive region grows；
Fig. 3 is the result that adaptive region is grown and Morphological scale-space is obtained：(a) be human brain MR original images, (b) for pair
The hippocampus clip image answered, (c) is adaptive region growth result, and (d) is the border of connected region, and (e) is connected region
Convex polygon profile；
Fig. 4 is compared for the grey level histogram between Brain MR Image and hippocampus clip image：(a) it is human brain MR original images
Grey level histogram, (b) is the grey level histogram of correspondence hippocampus clip image；
Fig. 5 is Li methods and the result for improving level set algorithm processing：(a) it is human brain MR original images, (b) is Li methods
Hippocampus contours extract result, (c) is that, using the result improved after level set algorithm processing, (d) is that doctor delineates result by hand；
Fig. 6 is segmentation result of the improved model to other tissues.
Embodiment
The following is specific embodiment of the invention and with reference to accompanying drawing, technical scheme is further described,
But the present invention is not limited to these embodiments.
As depicted in figs. 1 and 2, a kind of level set hippocampus image partition method initialized based on region growing, including under
Row step：
Step S1：Such as Fig. 3 (a) input original images, such as Fig. 3 (b) chooses one from the original image of whole hippocampal formations
Seed point, such as Fig. 4 are partitioned into the clip image for including whole hippocampal formation regions by seed point, and to the ash in seed point region
Degree average is calculated；
Step S2：Seed point and the similitude of the neighborhood territory pixel of seed point 3 × 3 are calculated using criterion is grown, with identical
Or the pixel of similar gray-value is referred to the region where sub-pixel, and adaptive region is reused to the pixel newly obtained
Grow criterion to calculate, until not new pixel is received, algorithm terminates, growth result such as Fig. 3 (c)；
Step S3：The interference that the bridge joint of respective pixel point and branch are present is removed with morphological operation, obtains orderly
Hippocampus contour line.
Step S4：On the basis of obtaining orderly hippocampus contour line, to incorporate global gaussian probability information, using improved
Level set algorithm realizes the segmentation of hippocampus, and driving contour curve stops to gtoal setting and in hippocampus boundary.
It is in step sl, described in the above-mentioned level set hippocampus image partition method initialized based on region growing
The calculation formula of gray average is in seed point region：
In formula,To calculate obtained seed region gray average；(x, y) is the seed point coordinates manually chosen；I
(i, j) is the gray value at pixel (i, j) place；R is seed region.
Generally, the growth criterion of design section growth algorithm can be directly carried out with the gray average of sub-pixel.
But, as n=1 (n represents the number of pixel in seed region), hippocampus is because the presence of noise spot, the knot of growth
Just very likely there is certain deviation in fruit.At this time, it may be necessary to the gray average in sub-pixel neighborhood be obtained, with it come substitute species
Gray average in subregion, be thus avoided as much as seed point falsely drops the shadow produced with noise spot to growth result
Ring.
In the above-mentioned level set hippocampus image partition method initialized based on region growing, in step sl, from cutting
The standard deviation of pixel grey scale in image-region is cut as the threshold values of growth, the gray average in clip image regionAnd variance
σGlobalCalculation formula be：
In formula, M × N is the size of clip image.
Just can adaptively it be adjusted according to the gray feature of image in itself using such method growth threshold values, can
The blindness of artificial selection threshold values is effectively prevented from, the adaptivity of region growing can be strengthened.
In the above-mentioned level set hippocampus image partition method initialized based on region growing, the meter of described growth criterion
Calculating formula is：
In formula, ξ is standard deviation sigmaGlobalWeight.
In the above-mentioned level set hippocampus image partition method initialized based on region growing, step S3 includes following sub-step
Suddenly：
1. hole region in region growing result is filled by S3；
2. S3 the target in image is marked；
3. S3 implements opening operation to marked region；
4. S3 extracts the connected region where image center；
5. S3 obtains the profile of connected region, such as Fig. 3 (d)；
6. S3 extracts the convex polygon profile of connected region, such as Fig. 3 (e).
In the above-mentioned level set hippocampus image partition method initialized based on region growing, step S4 includes following sub-step
Suddenly：
1. S4 sets up level set movements equation：
In formula, φ is level set function, and I (x) is initial pictures, and λ, ν, μ, τ are respectively length, area, canonical
The weight coefficient of item and global Gauss Distribution Fitting energy term, div is bent curvature of a curve, δεAnd H (g)ε(g) it is respectively Dirac letters
δ (z) and Heaviside function H (z) approximate expression are counted, its calculation formula is：
G is edge indicator function, and its calculation formula is：
In formula,To represent image gradient, GσThe Gaussian convolution function for being σ for standard deviation.
e1、e2For global Gauss Distribution Fitting energy term, its calculation formula is：
In formula, uiFor the global average of gray level image, σiPoor for the global criteria of gray level image, its calculation formula is：
2. S4 initializes level set function φ (x, y, t)=0 using the result of criterion algorithm and morphological operation is grown；
S4 3. initiation parameter μ, λ, ν, τ, ε, σ, Δ t；
4. S4 updates e1、e2；
5. S4 utilizes finite difference calculus, updates level set function φ；
6. whether determined level collection evolution curve meets convergence criterion to S4, if not provided, and go to step S4 4. continue meter
Calculate, until meeting end condition.
In order to quantitatively compare each model segmentation precision, using false negative rate (False Negative Ratio,
FNR), segmentation error rate (Ratio of Segmentation Error, RSE), Dice likeness coefficients (Dice
Similarity Coefficient, DSC) segmentation precision of each method is weighed, calculation formula is as follows：
In formula, N (*) represents the number of pixel in closed area.FNR, RSE value more connect closer to 0, DSC value
Nearly 1, represent that the segmentation precision of image is higher.
Table 1 provides each algorithm to the quantitative evaluation of hippocampus image segmentation result in Fig. 5, by quantitative comparison, changes
It is highest to enter DSC value of the model obtained in all images, illustrates that the method for the invention is better than Li algorithms in precision.
The ratio of precision of each algorithm of table 1 segmentation hippocampus image is compared with (image is from left to right numbered)
Li methods are shown in Table 2 with the improved model time required in hippocampus contours extract and iterations.Calculated due to each
Time needed for method initialization is identical (about 5s), therefore is not taken statistics in table 2.As can be seen from the table, needed for two kinds of algorithms
Time and iterations difference less, almost can be ignored, and doctor delineates hippocampus profile on high-resolution thin layer
Longer (average in 30s or so) is then taken, is split less efficient.Therefore the method for the invention has necessarily in precision and efficiency
Advantage, be easy to implement clinical application in real time.
The sliced time of each model and iterations in the Fig. 4-9 of table 2
Model of the embodiment of the present invention is as shown in Figure 6 to the segmentation result to other tissues.
Specific embodiment described herein is only to spirit explanation for example of the invention.Technology neck belonging to of the invention
The technical staff in domain can be made various modifications or supplement to described specific embodiment or be replaced using similar mode
Generation, but without departing from the spiritual of the present invention or surmount scope defined in appended claims.
Although there is used herein more term, being not precluded from the possibility using other terms.Using these terms only
Merely to more easily describing and explaining the essence of the present invention；Being construed as any additional limitation is all and this
What spirit was disagreed.
Claims (6)
1. a kind of level set hippocampus image partition method initialized based on region growing, it is characterised in that comprise the following steps：
Step S1：Original image is inputted, a seed point is chosen from the original image of whole hippocampal formations, is partitioned into by seed point
The clip image in whole hippocampal formation regions is included, and the gray average in seed point region is calculated；
Step S2：Seed point and the similitude of the neighborhood territory pixel of seed point 3 × 3 are calculated using criterion is grown, with identical or phase
The pixel of ashy angle value is referred to the region where sub-pixel, and the pixel recycling adaptive region growth to newly obtaining
Criterion is calculated, and until not new pixel is received, algorithm terminates；
Step S3：The interference that the bridge joint of respective pixel point and branch are present is removed with morphological operation, orderly hippocampus is obtained
Contour line.
Step S4：On the basis of obtaining orderly hippocampus contour line, to incorporate global gaussian probability information, improved level is utilized
Set algorithm realizes the segmentation of hippocampus, and driving contour curve stops to gtoal setting and in hippocampus boundary.
2. the level set hippocampus image partition method according to claim 1 initialized based on region growing, its feature is existed
In in step sl, the calculation formula of gray average is in described seed point region：
In formula,To calculate obtained seed region gray average；(x, y) is the seed point coordinates manually chosen；I(i,j)
For the gray value at pixel (i, j) place；R is seed region.
3. the level set hippocampus image partition method according to claim 1 initialized based on region growing, its feature is existed
In, in step sl, from pixel grey scale in clip image region standard deviation as growth threshold values, in clip image region
Gray average X and variances sigmaGlobalCalculation formula be：
In formula, M × N is the size of clip image.
4. the level set hippocampus image partition method according to claim 3 initialized based on region growing, its feature is existed
In the calculation formula of described growth criterion is：
In formula, ξ is standard deviation sigmaGlobalWeight.
5. the level set hippocampus image partition method according to claim 1 initialized based on region growing, its feature is existed
In step S3 includes substep：
1. hole region in region growing result is filled by S3；
2. S3 the target in image is marked；
3. S3 implements opening operation to marked region；
4. S3 extracts the connected region where image center；
5. S3 obtains the profile of connected region；
6. S3 extracts the convex polygon profile of connected region.
6. the level set hippocampus image partition method according to claim 1 initialized based on region growing, its feature is existed
In step S4 includes substep：
1. S4 sets up level set movements equation：
In formula, φ is level set function, and I (x) is initial pictures, λ, ν, μ, τ be respectively length, area, regular terms and
The weight coefficient of global Gauss Distribution Fitting energy term, div is bent curvature of a curve, δεAnd H (g)ε(g) it is respectively Dirac functions δ
(z) with Heaviside function H (z) approximate expression, its calculation formula is：
G is edge indicator function, and its calculation formula is：
In formula, ▽ I are expression image gradient, GσThe Gaussian convolution function for being σ for standard deviation.
e1、e2For global Gauss Distribution Fitting energy term, its calculation formula is：
In formula, uiFor the global average of gray level image, σiPoor for the global criteria of gray level image, its calculation formula is：
2. S4 initializes level set function φ (x, y, t)=0 using the result of criterion algorithm and morphological operation is grown；
S4 3. initiation parameter μ, λ, ν, τ, ε, σ, Δ t；
4. S4 updates e1、e2；
5. S4 utilizes finite difference calculus, updates level set function φ；
6. whether determined level collection evolution curve meets convergence criterion to S4, if do not have, and go to step S4 4. continue calculating, until
Meet end condition.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710225569.1A CN107016683A (en) | 2017-04-07 | 2017-04-07 | The level set hippocampus image partition method initialized based on region growing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710225569.1A CN107016683A (en) | 2017-04-07 | 2017-04-07 | The level set hippocampus image partition method initialized based on region growing |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107016683A true CN107016683A (en) | 2017-08-04 |
Family
ID=59445513
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710225569.1A Pending CN107016683A (en) | 2017-04-07 | 2017-04-07 | The level set hippocampus image partition method initialized based on region growing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107016683A (en) |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107481252A (en) * | 2017-08-24 | 2017-12-15 | 上海术理智能科技有限公司 | Dividing method, device, medium and the electronic equipment of medical image |
| CN107705303A (en) * | 2017-10-16 | 2018-02-16 | 长沙乐成医疗科技有限公司 | The dividing method of blood vessel on a kind of medical image |
| CN107767388A (en) * | 2017-11-01 | 2018-03-06 | 重庆邮电大学 | A kind of image partition method of combination cloud model and level set |
| CN108520525A (en) * | 2018-04-12 | 2018-09-11 | 重庆理工大学 | A kind of spinal cord dividing method based on convex constraint seed region growth |
| 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 |
| CN109509203A (en) * | 2018-10-17 | 2019-03-22 | 哈尔滨理工大学 | A kind of semi-automatic brain image dividing method |
| CN110599505A (en) * | 2019-09-17 | 2019-12-20 | 上海微创医疗器械（集团）有限公司 | Organ image segmentation method and device, electronic equipment and storage medium |
| WO2020029064A1 (en) * | 2018-08-07 | 2020-02-13 | 温州医科大学 | Optical coherence tomographic image processing method |
| CN111127479A (en) * | 2019-12-17 | 2020-05-08 | 昆明理工大学 | Level set image segmentation method based on curve area |
| CN111898600A (en) * | 2020-07-10 | 2020-11-06 | 浙江大华技术股份有限公司 | Character outline extraction method and device, storage medium and electronic device |
| CN111986216A (en) * | 2020-09-02 | 2020-11-24 | 长春工业大学 | RSG liver CT image interactive segmentation algorithm based on neural network improvement |
| US10970604B2 (en) | 2018-09-27 | 2021-04-06 | Industrial Technology Research Institute | Fusion-based classifier, classification method, and classification system |
| CN115082468A (en) * | 2022-08-22 | 2022-09-20 | 江苏思伽循环科技有限公司 | Electrode material separation control method and system in power battery recovery process |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101982835A (en) * | 2010-11-12 | 2011-03-02 | 西安电子科技大学 | Level set method for edge detection of SAR images of airport roads |
| CN104599270A (en) * | 2015-01-18 | 2015-05-06 | 北京工业大学 | Breast neoplasms ultrasonic image segmentation method based on improved level set algorithm |
| CN105321184A (en) * | 2015-11-26 | 2016-02-10 | 北京交通大学 | Improved edge level set-based method and system for segmenting noisy image |
| CN106447682A (en) * | 2016-08-29 | 2017-02-22 | 天津大学 | Automatic segmentation method for breast MRI focus based on Inter-frame correlation |
-
2017
- 2017-04-07 CN CN201710225569.1A patent/CN107016683A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101982835A (en) * | 2010-11-12 | 2011-03-02 | 西安电子科技大学 | Level set method for edge detection of SAR images of airport roads |
| CN104599270A (en) * | 2015-01-18 | 2015-05-06 | 北京工业大学 | Breast neoplasms ultrasonic image segmentation method based on improved level set algorithm |
| CN105321184A (en) * | 2015-11-26 | 2016-02-10 | 北京交通大学 | Improved edge level set-based method and system for segmenting noisy image |
| CN106447682A (en) * | 2016-08-29 | 2017-02-22 | 天津大学 | Automatic segmentation method for breast MRI focus based on Inter-frame correlation |
Non-Patent Citations (2)
| Title |
|---|
| 付英杰: "结合区域生长和水平集的海马轮廓提取和三维重建", 《中国优秀硕士学位论文全文数据库-信息科技辑》 * |
| 王辉: "图像分割的最优化和水平集方法研究", 《中国博士学位论文全文数据库-信息科技辑》 * |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107481252A (en) * | 2017-08-24 | 2017-12-15 | 上海术理智能科技有限公司 | Dividing method, device, medium and the electronic equipment of medical image |
| CN107705303A (en) * | 2017-10-16 | 2018-02-16 | 长沙乐成医疗科技有限公司 | The dividing method of blood vessel on a kind of medical image |
| CN107767388B (en) * | 2017-11-01 | 2021-02-09 | 重庆邮电大学 | Image segmentation method combining cloud model and level set |
| CN107767388A (en) * | 2017-11-01 | 2018-03-06 | 重庆邮电大学 | A kind of image partition method of combination cloud model and level set |
| CN108520525A (en) * | 2018-04-12 | 2018-09-11 | 重庆理工大学 | A kind of spinal cord dividing method based on convex constraint seed region growth |
| CN108520525B (en) * | 2018-04-12 | 2021-11-02 | 重庆理工大学 | Spinal cord segmentation method based on convex constraint seed region growth |
| 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 |
| WO2020029064A1 (en) * | 2018-08-07 | 2020-02-13 | 温州医科大学 | Optical coherence tomographic image processing method |
| US10970604B2 (en) | 2018-09-27 | 2021-04-06 | Industrial Technology Research Institute | Fusion-based classifier, classification method, and classification system |
| CN109509203A (en) * | 2018-10-17 | 2019-03-22 | 哈尔滨理工大学 | A kind of semi-automatic brain image dividing method |
| CN110599505A (en) * | 2019-09-17 | 2019-12-20 | 上海微创医疗器械（集团）有限公司 | Organ image segmentation method and device, electronic equipment and storage medium |
| CN111127479A (en) * | 2019-12-17 | 2020-05-08 | 昆明理工大学 | Level set image segmentation method based on curve area |
| CN111898600A (en) * | 2020-07-10 | 2020-11-06 | 浙江大华技术股份有限公司 | Character outline extraction method and device, storage medium and electronic device |
| CN111986216A (en) * | 2020-09-02 | 2020-11-24 | 长春工业大学 | RSG liver CT image interactive segmentation algorithm based on neural network improvement |
| CN111986216B (en) * | 2020-09-02 | 2024-01-12 | 无锡学院 | RSG liver CT image interactive segmentation algorithm based on neural network improvement |
| CN115082468A (en) * | 2022-08-22 | 2022-09-20 | 江苏思伽循环科技有限公司 | Electrode material separation control method and system in power battery recovery process |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107016683A (en) | The level set hippocampus image partition method initialized based on region growing | |
| CN107230206B (en) | Multi-mode data-based 3D pulmonary nodule segmentation method for hyper-voxel sequence lung image | |
| CN105957063B (en) | CT image liver segmentation method and system based on multiple dimensioned weighting similarity measure | |
| US9445744B2 (en) | Methods, systems, and devices for spine centrum extraction and intervertebral disk dividing | |
| CN110335358B (en) | Personalized deformation method of deformable digital human anatomy model | |
| CN107146228B (en) | A kind of super voxel generation method of brain magnetic resonance image based on priori knowledge | |
| CN106408569A (en) | Brain MRI (magnetic resonance image) segmentation method based on improved fuzzy C-means clustering algorithm | |
| CN101576997B (en) | Abdominal organ segmentation method based on secondary three-dimensional region growth | |
| CN110120048B (en) | Three-dimensional brain tumor image segmentation method combining improved U-Net and CMF | |
| Kanade et al. | Brain tumor detection using MRI images | |
| CN107680107B (en) | Automatic segmentation method of diffusion tensor magnetic resonance image based on multiple maps | |
| CN103942780B (en) | Based on the thalamus and its minor structure dividing method that improve fuzzy connectedness algorithm | |
| CN107563996B (en) | A kind of novel discus nervi optici dividing method and system | |
| CN109727235A (en) | A kind of automatic delineation algorithms of organ based on deep learning | |
| CN109671054A (en) | The non-formaldehyde finishing method of multi-modal brain tumor MRI | |
| CN106898000B (en) | Automatic segmentation method for magnetic resonance imaging brain gray matter nuclei | |
| CN108961278B (en) | Method and system for abdominal wall muscle segmentation based on image data | |
| Kumar et al. | Image processing in diabetic related causes | |
| CN109685804A (en) | A kind of multichannel head NMR zeugmatographic imaging tissue segmentation methods | |
| CN104318565A (en) | Interactive method for retinal vessel segmentation based on bidirectional region growing of constant-gradient distance | |
| CN104463885B (en) | A kind of Multiple Sclerosis lesions region segmentation method | |
| CN106919950B (en) | The brain MR image segmentation of probability density weighting geodesic distance | |
| CN104915989B (en) | Blood vessel three-dimensional dividing method based on CT images | |
| CN106446805A (en) | Segmentation method and system for optic cup in eye ground photo | |
| CN103985113B (en) | Tongue is as dividing method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170804 |