CN114972764B - Multi-atlas segmentation method based on feature clustering - Google Patents
Multi-atlas segmentation method based on feature clustering Download PDFInfo
- Publication number
- CN114972764B CN114972764B CN202210914811.7A CN202210914811A CN114972764B CN 114972764 B CN114972764 B CN 114972764B CN 202210914811 A CN202210914811 A CN 202210914811A CN 114972764 B CN114972764 B CN 114972764B
- Authority
- CN
- China
- Prior art keywords
- image
- map
- volume
- segmentation
- library
- 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
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
- G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/762—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using clustering, e.g. of similar faces in social networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/03—Recognition of patterns in medical or anatomical images
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computing Systems (AREA)
- Databases & Information Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
The invention relates to the field of medical image processing, in particular to a multi-atlas segmentation method based on feature clustering, which comprises the following steps: performing delineation, naming and volume calculation on the medical image, constructing a volume characteristic value list according to a calculation result and adjusting to form a general map library; mapping the volume characteristic value into a normalized data interval, classifying the map library according to the volume characteristic by using a clustering algorithm, and randomly initializing n clustering central points; calculating the distance from the volume characteristic value to the central point, determining the cluster according to the distance, and dividing the graph library into n sub-graph libraries; and calculating the distance between the image to be segmented and the image volume characteristic value of each map library, selecting a map with a closer distance as a map required by the image segmentation, and performing multi-map segmentation in the corresponding sub-library to obtain a segmentation result. Compared with the existing atlas segmentation method, the improved multi-feature atlas segmentation method has better effect, and improves the efficiency and the precision of the segmentation of the large volume interesting region CT image.
Description
Technical Field
The invention relates to the technical field of medical image processing, in particular to a multi-atlas segmentation method based on feature clustering.
Background
The medical image segmentation aims to distinguish different regions with specific significance, any two regions are not crossed, and each region meets the consistency of attributes, so that the medical image segmentation has very important significance in the fields of three-dimensional visualization, pathological analysis, clinical diagnosis, medical intervention and the like.
The atlas segmentation method based on the deformation algorithm is an important image segmentation method. The method takes the boundary of an interest region as an optimization target, accurately delineates a well-defined outline in a spectrogram image, maps the accurately-delineated outline into an image to be segmented through a deformation field and forms a new outline. The normal tissue region-of-interest image segmentation of the CT image is an important application of the algorithm. Among them, multi-atlas segmentation (MAS) is one of the most widely used image segmentation techniques in the field of medical image segmentation. And the multi-map segmentation allocates a segmentation label to each voxel of the image to be segmented by utilizing the corresponding relation of the similarity information of the map image and the image to be segmented. The multi-atlas segmentation has advantages over other segmentation methods because the atlas image used contains a priori knowledge of the region of interest that has been delineated by an expert in advance.
The existing multi-atlas segmentation method selects a group of CT images with the highest similarity measure with the image to be segmented from an image library, and the similarity measure comprises the sum of squares of errors between the images, related and mutual information, anatomical structures and the like. Although a suitable matching atlas can be found, when searching in a huge atlas database, the similarity measure between each image in the atlas database and the image to be segmented needs to be calculated, so that a large amount of computer resources need to be consumed, difficulty is brought to searching, and the whole segmentation process is more complicated. More importantly, the conventional Atlas image selection method mainly takes the similarity of the CT value (gray level) between the Atlas image and the image to be segmented as a selection standard, rather than the similarity of an interested area as a standard, and the similarity of the image outside the interested area can interfere with the result of Atlas image selection.
The present invention seeks to address these and other needs in the art.
Disclosure of Invention
In order to solve at least one technical problem mentioned in the background technology, the invention provides a new improved graph spectrum search algorithm-a multi-feature graph spectrum library search algorithm by combining the respective advantages of a similarity measure search method and a contour feature search method. Experiments prove that the improved multi-feature map segmentation method has better effect than the existing map segmentation method, and improves the efficiency and the precision of the segmentation of the large-volume interested region CT image.
A multi-atlas segmentation method based on feature clustering comprises
Performing delineation, naming and volume calculation on the medical image, constructing a volume characteristic value list according to a calculation result and adjusting to form a general map library;
mapping the volume characteristic values to normalized data intervals to form a list, classifying the map library according to the volume characteristics by using a clustering algorithm, and randomly initializing n clustering central points;
calculating the distance from the volume characteristic value to the central point, determining the clustering according to the distance, and dividing the spectrum library into n sub-spectrum libraries according to the clustering result;
and calculating the distance between the image to be segmented and the image volume characteristic value of each map library, selecting a map with a closer distance as a map required by the image segmentation, and performing multi-map segmentation in the corresponding sub-library to obtain a segmentation result.
Further, the multi-atlas segmentation method based on feature clustering specifically comprises the following steps:
s1, delineating a CT image normal tissue region in a map library;
s2, naming the tissue contour obtained by drawing;
s3, calculating the normal tissue contour volume of the CT image;
s4, constructing a profile volume characteristic value list corresponding to the map according to the volume calculation result;
s5, adjusting the image of the map library according to the organ volume data to form a general map library;
s6, mapping the volume value of the same contour sample in the spectrum library to a normalized data interval to form a list, and keeping the list as a volume characteristic value of the image of the spectrum library;
s7, classifying the map library according to the volume characteristics by using a clustering algorithm, and randomly initializing n clustering central points, wherein each central point uses a volume characteristic value (V) 1 ,V 2 ,V 3 …) as different dimensions;
s8, calculating the distance from the volume characteristic value of each contour of a certain group of images in the atlas database to a central point, determining the cluster to which the contour belongs according to the distance, and dividing the atlas database into n sub-atlas databases according to the clustering result;
s9, calculating the distance between the image to be segmented and the image volume characteristic value of each map library, and selecting a map with a closer distance as a map required by image segmentation;
and S10, performing multi-map segmentation in the corresponding sub-library to obtain a final segmentation result.
Further, in step S1, the CT image includes a Dicom format.
Further, in step S2, when naming the delineated tissue contour, the same contour name is used for different CT images.
Further, in step S5, the organ volume data includes maximum and minimum volumes of the organ volume and frequency of occurrence thereof.
Further, in step S5, the adjusting the gallery image includes adding and/or gallery images.
Further, in step S6, a normalization algorithm is applied when the volume values of the same contour sample in the map library are mapped to a normalized data interval to form a list.
Further, in step S9, when the atlas database is selected, the contour volume of the image to be segmented may be estimated in advance, and a corresponding atlas database is determined.
Further, in step S9, the selecting a map with a closer distance as the map required to be used for the image segmentation means that the closest 3-5 map is selected as the map required to be used for the image segmentation according to the proximity.
The multi-feature map library search algorithm provided by the invention uses the volume characteristic value as the basis for determining the map sub-library, changes the mechanism that the traditional map library search algorithm uses the similarity measure as the best matching map search algorithm, and avoids the problem that the deformation algorithm can not be accurately deformed due to the fact that the overall similarity of the images is close but the local volume difference of the outline of the region of interest is too large, so that more accurate deformation results can be obtained.
According to the method, the large number of the map libraries are divided into n sub-libraries, and after the contour volume of the map to be segmented is pre-estimated, the segmentation is completed only by using the sub-map libraries, so that the time for searching the best matching map by an algorithm is shortened, and the image segmentation efficiency is improved.
An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor executes the program to perform the feature clustering-based multi-atlas segmentation method.
A computer-readable storage medium having stored thereon computer instructions for causing the computer to execute the aforementioned feature clustering-based multi-atlas segmentation method.
The preferred conditions described above may be combined with each other to arrive at a specific embodiment, based on general knowledge in the art.
The beneficial effects of the invention are as follows:
1) When an image of the atlas database is selected, each organ is selected according to the volume value, the volume of each organ of the atlas in the database is ensured to have approximate frequency distribution, and the aim is to ensure that a deformation algorithm can more effectively calculate a segmentation result in the next segmentation process; the existing method does not perform similar treatment on the selected atlas;
2) Clustering and dividing the atlas database by using the volume characteristic values, wherein atlas samples in the divided atlas have approximate volume characteristic values and have reference values for dividing the image; the existing method does not divide a map library, the volume characteristic values of map samples are dispersed, the best matching map is often difficult to find during searching, and the deformation result is poor;
3) Pre-classifying the image to be segmented according to the volume characteristic of each contour of the image to be segmented, calculating the distance from the image to each sub-graph library clustering center according to the estimated volume characteristic value, and then selecting the sub-graph library with the minimum distance as a search graph library; the existing method does not pre-screen the image to be segmented, which often results in poor segmentation effect;
4) After clustering and dividing the spectrum libraries, the number of each sub-spectrum library is reduced, and when a new segmentation task is executed, the best matching spectrum is only required to be searched in a small range, so that the searching efficiency is high, and the calculation cost is low; the existing method uses a complete spectrum library, and the segmentation needs to be searched in the range of the full spectrum library, so that the searching efficiency is low, and the calculation cost is high;
5) The invention can also use the volume characteristic value to determine the best matching map, and the best matching map is determined by calculating the distance between the image to be segmented and the image volume characteristic value of each map library. The method has the advantages of short calculation time, low calculation cost and high segmentation accuracy; the existing method determines the best matching map by using similarity measurement, and determines the best matching map by calculating the similarity measurement between the image to be segmented and each map library image.
The invention adopts the technical scheme for achieving the purpose, makes up the defects of the prior art, and has reasonable design and convenient operation.
Drawings
The accompanying drawings, which are included to provide a further understanding of the disclosure and are incorporated in and constitute a part of this specification, illustrate embodiments of the disclosure and together with the description serve to explain the principles of the disclosure.
FIG. 1 is a general flow chart of the segmentation method of the present invention;
FIG. 2 is a schematic diagram of the clustering method described in example 2;
FIG. 3 is a flowchart of a clustering method according to embodiment 2;
fig. 4 shows the segmentation method according to the present invention and the prior art segmentation method in the bladder and rectum segmentation.
Detailed Description
Those skilled in the art can appropriately substitute and/or modify the process parameters to implement the present disclosure, but it is specifically noted that all similar substitutes and/or modifications will be apparent to those skilled in the art and are deemed to be included in the present invention. While the invention has been described in terms of preferred embodiments, it will be apparent to those skilled in the art that the technology can be practiced and applied by modifying or appropriately combining the embodiments described herein without departing from the spirit and scope of the invention.
It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the disclosure herein. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.
It is to be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of the present application. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, and it should be understood that when the terms "comprises" and/or "comprising" are used in this specification, they specify the presence of stated features, steps, operations, devices, components, and/or combinations thereof, unless the context clearly indicates otherwise.
The present invention is described in detail below.
Example 1:
as shown in fig. 1, a multi-atlas segmentation method based on feature clustering is provided, which first includes the creation of an atlas database, and specifically includes the following steps.
1. Creating a atlas database containing 100-150 CT images, and assigning uniform name labels to the contours in each CT image, such as: "blade", "resume", "External", "Femal _ Head _ Left", "Femal _ Head _ Right", and the like. Checking the delineation accuracy of each CT image, and ensuring the consistency of contour delineation in the atlas database.
2. Medical image processing software is used for counting the volumes of all CT image profiles, and the maximum value and the minimum value of all profile volumes in the image library are observed, so that all profile volume ranges of the selected image library can cover the profile volume of people, and the frequency distribution of the atlas in all profile volume ranges is uniform. For insufficient coverage of the contour volume value range, the database should be searched for new spectral library additions. For a selected atlas database but with high repeatability of the contour volume value, repeat atlases should be removed from the atlas database.
TABLE 1 atlas database image/silhouette volume
| Original | Contour 1 (cm) 3 ) | Contour 2 (cm) 3 ) | Contour 3 (cm) 3 ) | Contour 4 (cm) 3 ) |
| Image 1 | 74.73 | 52.89 | 136.75 | 31776.14 |
| Image 2 | 157.52 | 107.02 | 101.09 | 22976.85 |
| … | … | … | … | … |
| |
430.22 | 36.25 | 85.44 | 19524.16 |
TABLE 2 volume eigenvalues for atlas database image/contour normalization processing
| Normalization | Contour 1 (cm) 3 ) | Contour 2 (cm) 3 ) | Contour 3 (cm) 3 ) | Contour 4 (cm) 3 ) |
| Image 1 | 0.17 | 0.49 | 1.00 | 1.00 |
| Image 2 | 0.37 | 1.00 | 0.74 | 0.72 |
| … | … | … | … | … |
| |
1 | 0.34 | 0.62 | 0.61 |
3. And calculating the volume characteristic value of each image normalization processing in the atlas database according to the selected atlas to serve as the volume characteristic of the atlas.
4. As shown in fig. 2 and 3, the images in the atlas database are clustered by different volume features using a clustering algorithm. For example, volume features of 'Blader', 'Rectum' and 'External' are selected to form a three-dimensional array, 3-4 clustering centers are determined, and 3-4 sub-atlas databases are determined through calculation after multiple iterations.
5. The cluster center of each sub-graph library is recorded.
Example 2:
on the basis of the preceding embodiment, sub-graph spectral segmentation is used.
1. And (3) using medical image processing software, estimating the approximate volume of each contour to be drawn according to an empirical formula and the number of CT layers occupied by each contour to be drawn, and carrying out normalized calculation on the volume.
2. And calculating the distance between the image to be segmented and the clustering center of each sub-spectrum library, and selecting the library with the closest distance as the sub-spectrum library required to be used for segmenting the image.
3. Respectively recording and selecting required interested areas on the image to be segmented by using a multi-atlas segmentation method, searching 3-5 atlases most similar to the image to be segmented in a library, fusing the 3-5 atlases, and then performing deformation registration on the image to be segmented to obtain a segmentation result.
Verification example:
the accuracy of segmentation of pelvic organ images by using the prior art method and the method provided by the previous embodiment of the application is verified. The experiment was performed on a Raystation treatment planning System version 9.2 platform test, including an atlas library of 100 CT's, using a multi-atlas segmentation method as a baseline method. 30 test set cases outside the atlas database are randomly selected, and the segmentation accuracy of the bladder and rectum is respectively tested.
The spectral library containing 100 cases was divided into 4 sub-libraries using a cluster segmentation method. After clustering is carried out according to the volume characteristics and normalized clustering is carried out, the number of the chart libraries contained in each sub-library is respectively as follows: 32, 30, 21, 17. Normalization adjusts the distribution of the sub-library, so that the image to be segmented has higher probability to be distributed to a better atlas library.
The segmentation results were evaluated by a similarity coefficient (DSC), defined as:
wherein V ref Exact profile volume, V, sketched for an expert auto The contour volume is delineated for the segmentation method. The DCS is 0 to 1 in size, and when the DSC =1, the DCS and the DSC are completely overlapped to realize perfect segmentation; when DSC =0, it indicates that the two do not intersect each other, and there is no segmentation effect. Using SPSS 21 software to compare, analyze and delineate the difference of the results, testing the results to be in accordance with normal distribution, performing pairing t test and P test on different segmentation modes<The difference of 0.05 is statistically significant, and the segmentation effect is shown in table 3 and fig. 4.
TABLE 3 segmentation Effect
As can be seen from table 3 and fig. 4, the normalized multi-feature segmentation method has a much higher bladder segmentation effect than the existing multi-map segmentation method (0.83 ± 0.08 vs 0.69 ± 0.15, p-straw 0.05), and has a much higher rectal segmentation effect than the existing multi-map segmentation method (0.70 ± 0.12 vs 0.56 ± 0.15, p-straw 0.05), and has significant differences from the existing segmentation methods. In terms of segmentation time, the existing segmentation method has a long automatic delineation time (6.3 +/-0.1 min) due to a large number of template cases (100 cases) in the atlas database, while the method provided by the embodiment of the application takes much less time (2.6 +/-0.5 min) than the existing segmentation method because the template cases in the atlas database are simplified.
Therefore, the application provides a multi-atlas segmentation method based on feature clustering, which comprises the steps of firstly carrying out delineation, naming and volume calculation on a medical image, constructing a volume feature value list according to a calculation result and adjusting to form a total atlas database; secondly, mapping the volume characteristic values to normalized data intervals to form a list, classifying the map library by using a clustering algorithm according to the volume characteristics, and randomly initializing n clustering central points; calculating the distance from the volume characteristic value to the central point again, determining the clustering according to the distance, and dividing the spectrum library into n sub-spectrum libraries according to the clustering result; and finally, calculating the distance between the image to be segmented and the image volume characteristic value of each map library, selecting a map with a closer distance as a map required by the image segmentation, and performing multi-map segmentation in a corresponding sub-library to obtain a segmentation result. By dividing the map library with a large number into n sub-libraries and performing pre-estimation on the contour volume of the map to be segmented, the segmentation is completed only by using the sub-library, so that the time for searching the best matching map by an algorithm is shortened, and the image segmentation efficiency is improved.
Example 3:
an electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor executes the program to perform the feature clustering-based multi-atlas segmentation method.
Example 4:
a computer-readable storage medium having stored thereon computer instructions for causing the computer to execute the aforementioned feature clustering-based multi-atlas segmentation method.
The substantial content of the technical solutions of the present invention or the core content part contributing to the prior art may be embodied in the form of a software product, which may be stored in a computer memory or a storage medium, and which, when implementing the present invention, further includes several instructions to make a computer device perform all or part of the steps of the method according to the technical solutions of the present invention, the memory or the storage medium includes a Random Access Memory (RAM), a memory, a Read Only Memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the technical field.
The above-mentioned devices are only exemplary, the units in the electronic equipment and the computer may be physically separated or not, and the parts of the units can be selected according to actual situations and needs to realize the essence of the technical solution of the present invention, so that those skilled in the art can understand and implement the technical solution of the present invention without creative efforts.
Conventional techniques in the above embodiments are known to those skilled in the art, and therefore, will not be described in detail herein.
The specific embodiments described herein are merely illustrative of the spirit of the invention. Various modifications or additions may be made to the described embodiments or alternatives may be employed by those skilled in the art without departing from the spirit or ambit of the invention as defined in the appended claims.
While the invention has been described in detail and with reference to specific examples thereof, it will be apparent to one skilled in the art that various changes and modifications can be made therein without departing from the spirit and scope thereof.
The above description is only a preferred embodiment of the present application and is not intended to limit the present application, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present application shall be included in the protection scope of the present application.
The invention is not the best known technology.
Claims (9)
1. A multi-atlas segmentation method based on feature clustering is characterized by comprising the following steps:
performing delineation, naming and volume calculation on the medical image, constructing a volume characteristic value list according to a calculation result and adjusting to form a general map library;
mapping the volume characteristic values to normalized data intervals to form a list, classifying the map library according to the volume characteristics by using a clustering algorithm, and randomly initializing n clustering central points;
calculating the distance from the volume characteristic value to the central point, determining the clustering according to the distance, and dividing the spectrum library into n sub-spectrum libraries according to the clustering result;
calculating the distance between the image to be segmented and the image volume characteristic value of each map library, selecting a map with a closer distance as a map required by the image segmentation, and performing multi-map segmentation in a corresponding sub-library to obtain a segmentation result;
the method comprises the following specific steps:
s1, delineating a CT image normal tissue region in a map library;
s2, naming the tissue contour obtained by drawing;
s3, calculating the normal tissue contour volume of the CT image;
s4, constructing a profile volume characteristic value list corresponding to the map according to the volume calculation result;
s5, adjusting the image of the map library according to the organ volume data to form a general map library;
s6, mapping the volume values of the same contour sample in the spectrum library to a normalized data interval to form a list, wherein the list is reserved as a volume characteristic value of the image of the spectrum library;
s7, classifying the atlas database according to the volume characteristics by using a clustering algorithm, and randomly initializing n clustering central points, wherein each central point uses a volume characteristic value (V) 1 ,V 2 ,V 3 …) as different dimensions;
s8, calculating the distance from the volume characteristic value of each contour of a certain group of images in the atlas database to a central point, determining the cluster to which the contour belongs according to the distance, and dividing the atlas database into n sub-atlas databases according to the clustering result;
s9, calculating the distance between the image to be segmented and the image volume characteristic value of each map library, and selecting a map with a closer distance as a map required by image segmentation;
and S10, performing multi-map segmentation in the corresponding sub-library to obtain a final segmentation result.
2. The method of claim 1, wherein: in step S1, the CT image includes a Dicom format.
3. The method of claim 1, wherein: in step S2, when naming the tissue contour obtained by delineation, the same contour name is used for different CT images.
4. A method according to any one of claims 1-3, characterized in that: in step S5, the organ volume data includes maximum and minimum volumes of the organ volume and the frequency of occurrence thereof.
5. A method according to any one of claims 1-3, characterized in that: in step S6, a normalization algorithm is applied when the volume values of the same contour sample in the spectrum library are mapped to a normalized data interval to form a list.
6. A method according to any one of claims 1-3, characterized in that: in step S9, when selecting the atlas database, the contour volume of the image to be segmented may be estimated in advance, and the corresponding atlas database is determined.
7. A method according to any one of claims 1-3, characterized in that: in step S9, the selecting a map with a relatively close distance as the map required to be used for the image segmentation means that the closest 3-5 map is selected as the map required to be used for the image segmentation according to the degree of closeness.
8. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein: the processor, when executing the program, performs the feature clustering based multi-atlas segmentation method of any of claims 1-7.
9. A computer-readable storage medium storing computer instructions, the computer-readable storage medium characterized in that: the instructions causing the computer to perform the method of feature clustering based multi-atlas segmentation of any of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210914811.7A CN114972764B (en) | 2022-08-01 | 2022-08-01 | Multi-atlas segmentation method based on feature clustering |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210914811.7A CN114972764B (en) | 2022-08-01 | 2022-08-01 | Multi-atlas segmentation method based on feature clustering |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114972764A CN114972764A (en) | 2022-08-30 |
| CN114972764B true CN114972764B (en) | 2022-11-18 |
Family
ID=82970113
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210914811.7A Active CN114972764B (en) | 2022-08-01 | 2022-08-01 | Multi-atlas segmentation method based on feature clustering |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114972764B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115687249B (en) * | 2022-12-30 | 2023-03-28 | 浙江大华技术股份有限公司 | Image gathering method and device, terminal and computer readable storage medium |
| CN116630341B (en) * | 2023-04-20 | 2023-12-01 | 广州方信医疗技术有限公司 | Pathological digital slide processing method and system |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2012164345A (en) * | 2006-10-10 | 2012-08-30 | Nikon Corp | Image classification program, image sorting device, and electronic camera |
| CN103426169A (en) * | 2013-07-26 | 2013-12-04 | 西安华海盈泰医疗信息技术有限公司 | Medical image segmentation algorithm |
| CN104637024A (en) * | 2013-11-14 | 2015-05-20 | 株式会社东芝 | Medical image data processing apparatus and method |
| CN109241628A (en) * | 2018-09-08 | 2019-01-18 | 西北工业大学 | Three-dimensional CAD model dividing method based on Graph Spectral Theory and cluster |
| CN112435263A (en) * | 2020-10-30 | 2021-03-02 | 苏州瑞派宁科技有限公司 | Medical image segmentation method, device, equipment, system and computer storage medium |
| CN114463248A (en) * | 2020-10-30 | 2022-05-10 | 国际商业机器公司 | Seed relabeling for seed-based segmentation of medical images |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107680107B (en) * | 2017-10-30 | 2020-09-08 | 西北工业大学 | Automatic segmentation method of diffusion tensor magnetic resonance image based on multiple maps |
| CN110378910B (en) * | 2019-06-28 | 2021-10-22 | 艾瑞迈迪医疗科技（北京）有限公司 | Abdominal cavity multi-organ segmentation method and device based on map fusion |
-
2022
- 2022-08-01 CN CN202210914811.7A patent/CN114972764B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2012164345A (en) * | 2006-10-10 | 2012-08-30 | Nikon Corp | Image classification program, image sorting device, and electronic camera |
| CN103426169A (en) * | 2013-07-26 | 2013-12-04 | 西安华海盈泰医疗信息技术有限公司 | Medical image segmentation algorithm |
| CN104637024A (en) * | 2013-11-14 | 2015-05-20 | 株式会社东芝 | Medical image data processing apparatus and method |
| CN109241628A (en) * | 2018-09-08 | 2019-01-18 | 西北工业大学 | Three-dimensional CAD model dividing method based on Graph Spectral Theory and cluster |
| CN112435263A (en) * | 2020-10-30 | 2021-03-02 | 苏州瑞派宁科技有限公司 | Medical image segmentation method, device, equipment, system and computer storage medium |
| CN114463248A (en) * | 2020-10-30 | 2022-05-10 | 国际商业机器公司 | Seed relabeling for seed-based segmentation of medical images |
Non-Patent Citations (2)
| Title |
|---|
| Agglomerative clustering of feature data for image segmentation;D.P. Mukherjee et al;《Proceedings. International Conference on Image Processing》;20021210;269-272 * |
| 基于无监督聚类的脑组织磁共振图像分割方法;张竞丹;《万方学位论文库》;20090921;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114972764A (en) | 2022-08-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114972764B (en) | Multi-atlas segmentation method based on feature clustering | |
| CN104933711B (en) | A kind of automatic fast partition method of cancer pathology image | |
| US9129390B2 (en) | Method of segmenting anatomic entities in 3D digital medical images | |
| CN111340828A (en) | Brain glioma segmentation based on cascaded convolutional neural networks | |
| CN109615642B (en) | Automatic organ-at-risk delineation method and device in radiotherapy plan | |
| WO2021203795A1 (en) | Pancreas ct automatic segmentation method based on saliency dense connection expansion convolutional network | |
| CN113706487A (en) | Multi-organ segmentation method based on self-supervision characteristic small sample learning | |
| Mahapatra et al. | Cardiac LV and RV segmentation using mutual context information | |
| Wu et al. | A supervoxel classification based method for multi-organ segmentation from abdominal ct images | |
| Dhara et al. | Differential geometry-based techniques for characterization of boundary roughness of pulmonary nodules in CT images | |
| CN111986101A (en) | Cerebrovascular map construction method | |
| CN112884820A (en) | Method, device and equipment for training initial image registration and neural network | |
| CN111008989B (en) | Airborne multispectral LIDAR three-dimensional segmentation method based on multivalued voxels | |
| CN115409879A (en) | Data processing method and device for image registration, storage medium and electronic equipment | |
| Cerrolaza et al. | Fetal skull segmentation in 3D ultrasound via structured geodesic random forest | |
| EP2006802A1 (en) | Method of constructing a grey value model and/or a geometric model of an anatomic entity in a 3D digital medical image | |
| CN114419047A (en) | Method, device, equipment and storage medium for determining blood vessel morphological characteristics | |
| CN113240680A (en) | Automatic colorectal cancer mesentery segmentation method based on deep learning technology | |
| CN103400385B (en) | The doubtful lymph node extracting method of stomach CT image that sparse dynamic integrity is selected | |
| US10573006B2 (en) | Selection of a subset of atlases from a set of candidate atlases | |
| CN109800820B (en) | Classification method based on ultrasound contrast image uniformity degree | |
| CN112529900A (en) | Method, device, terminal and storage medium for matching ROI in mammary gland image | |
| CN117244181A (en) | Dose analysis method and device based on radiotherapy risk organ outline | |
| CN1399763A (en) | 3-D shape measurements using statistical curvature analysis | |
| CN114821157A (en) | Multi-modal image classification method based on hybrid model network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |