CN113658187B - Medical image segmentation method, device and storage medium combined with anatomical priori - Google Patents
Medical image segmentation method, device and storage medium combined with anatomical priori Download PDFInfo
- Publication number
- CN113658187B CN113658187B CN202110842739.7A CN202110842739A CN113658187B CN 113658187 B CN113658187 B CN 113658187B CN 202110842739 A CN202110842739 A CN 202110842739A CN 113658187 B CN113658187 B CN 113658187B
- Authority
- CN
- China
- Prior art keywords
- segmentation
- data
- information
- medical image
- priori
- 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
- 238000003709 image segmentation Methods 0.000 title claims abstract description 67
- 238000000034 method Methods 0.000 title claims abstract description 39
- 230000011218 segmentation Effects 0.000 claims abstract description 146
- 210000003484 anatomy Anatomy 0.000 claims abstract description 65
- 238000012549 training Methods 0.000 claims abstract description 46
- 230000006870 function Effects 0.000 claims description 62
- 238000004590 computer program Methods 0.000 claims description 3
- 210000001519 tissue Anatomy 0.000 description 13
- 210000000056 organ Anatomy 0.000 description 10
- 238000003745 diagnosis Methods 0.000 description 3
- 238000005457 optimization Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000002591 computed tomography Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000002595 magnetic resonance imaging Methods 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 206010028980 Neoplasm Diseases 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 238000003759 clinical diagnosis Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000010191 image analysis Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 230000003902 lesion Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000007170 pathology Effects 0.000 description 1
- 238000004393 prognosis Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 210000004872 soft tissue Anatomy 0.000 description 1
- 230000007723 transport mechanism Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
-
- 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
- G06T2207/20081—Training; Learning
-
- 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
- G06T2207/20084—Artificial neural networks [ANN]
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The application discloses a medical image segmentation method, device and storage medium combined with anatomy priori. The medical image segmentation method combined with the anatomical prior comprises the following steps: obtaining segmentation training data; the segmentation training data comprises medical image data and segmentation label data; obtaining priori information of the anatomical structure according to the segmentation training data; calculating an overall loss function according to the medical image data, the segmentation label data and the anatomical structure priori information; and training the medical image data according to the total loss function and a preset image segmentation model to obtain target image data. The medical image segmentation method combined with the anatomical prior can improve the accuracy of medical image segmentation results.
Description
Technical Field
The present application relates to, but is not limited to, the field of computers, and in particular, to a method, apparatus, and storage medium for medical image segmentation in combination with anatomical priors.
Background
The medical image comprises human tissue and organ images obtained through various imaging modes such as X-ray, CT (Computer Tomography), MRI (Magnetic Resonance Imaging ), fundus color photograph and the like, and is an important basis for medical institutions to carry out diseases diagnosis, operation planning, prognosis evaluation, follow-up visit and the like, and the medical image processing is helpful for making the images more visual and clear, so that the diagnosis efficiency is improved.
In the field of medical image analysis, medical image segmentation plays an important role, and the purpose of medical image segmentation is to segment out a part with special meaning in a medical image, extract relevant characteristics, provide reliable basis for clinical diagnosis and pathology research, and assist doctors to make more accurate diagnosis. The image segmentation process segments the image into regions within which there are similar properties such as gray scale, color, texture, brightness, contrast, etc. Medical image segmentation is used clinically to study anatomical structures, measure tissue volumes, identify areas of tumors, lesions, and other abnormal tissue, and the like.
The medical image has higher complexity and lacks simple linear characteristics, and the accuracy of the medical image segmentation result is lower due to the influence of factors such as partial volume effect, gray level non-uniformity, artifacts, the proximity of gray levels among different soft tissues and the like.
Disclosure of Invention
The present application aims to solve at least one of the technical problems existing in the prior art. Therefore, the medical image segmentation method combined with the anatomical prior can improve the accuracy of medical image segmentation results.
An embodiment of a first aspect of the present application provides a medical image segmentation method in combination with anatomical prior, including: obtaining segmentation training data; the segmentation training data comprises medical image data and segmentation label data; obtaining priori information of the anatomical structure according to the segmentation training data; calculating an overall loss function according to the medical image data, the segmentation label data and the anatomical structure prior information; and training the medical image data according to the total loss function and a preset image segmentation model to obtain target image data.
The medical image segmentation method combined with the anatomical prior according to the embodiment of the application has at least the following technical effects: according to the medical image segmentation method combining the anatomical priori, anatomical structure priori information is obtained according to the segmentation training data, anatomical structure characteristics of tissue and organs are used as the priori information, the overall loss function is calculated, the anatomical structure priori information is utilized to guide the training and optimization of a medical image segmentation model, further, target image data after image segmentation are obtained, and accuracy of influencing a segmentation result is improved.
According to some embodiments of the application, the calculating the overall loss function from the medical image data, the segmentation label data, the anatomical structure prior information comprises: obtaining segmentation result data according to the medical image data and the preset image segmentation model; and obtaining the overall loss function according to the segmentation label data, the anatomical structure priori information and the segmentation result data.
According to some embodiments of the present application, the deriving the overall loss function from the segmentation label data, the anatomical prior information, segmentation result data comprises: calculating segmentation loss information according to the segmentation label data and the segmentation result data; calculating priori loss information according to the anatomical structure priori information and the segmentation result data; and obtaining the overall loss function according to the segmentation loss information and the prior loss information.
According to some embodiments of the application, the calculating a priori loss information from the anatomical structure a priori information, the segmentation result data includes: obtaining first sub priori loss information according to the segmentation result data and the segmentation label data; obtaining second sub priori loss information according to the segmentation result data; and obtaining the priori loss information according to the first sub priori loss information and the second sub priori loss information.
According to some embodiments of the present application, the obtaining the overall loss function according to the segmentation loss information and the prior loss information includes: acquiring loss weight information; and obtaining the overall loss function according to the segmentation loss information, the priori loss information and the loss weight information.
According to some embodiments of the application, the obtaining anatomical structure prior information from the segmentation training data includes: obtaining segmentation tag reconstruction data according to the segmentation tag data; and reconstructing data according to the segmentation labels to obtain prior information of the anatomical structure.
According to some embodiments of the application, the obtaining anatomical structure prior information from the segmentation tag reconstruction data includes: calculating a coding loss function according to the segmentation tag data and the segmentation tag reconstruction data; and obtaining the priori information of the anatomical structure according to the coding loss function.
An embodiment of a second aspect of the present application provides a medical image segmentation apparatus in combination with anatomical priors, comprising: the segmentation training data acquisition module is used for acquiring segmentation training data; the segmentation training data comprises medical image data and segmentation label data; the priori information acquisition module is used for acquiring anatomical structure priori information according to the segmentation training data; the overall loss function acquisition module is used for calculating an overall loss function according to the medical image data, the segmentation tag data and the anatomical structure priori information; and the target image data generation module is used for training the medical image data according to the overall loss function and a preset image segmentation model to obtain target image data.
An embodiment of a third aspect of the present application provides a medical image segmentation apparatus in combination with anatomical prior, comprising: a memory, a processor, and a computer program stored on the memory and executable on the processor, the processor implementing when executing the program: the above-described first aspect of the present application provides a medical image segmentation method in combination with anatomical prior.
According to an embodiment of the fourth aspect of the present application, a storage medium stores computer-executable instructions for: the medical image segmentation method in combination with anatomical prior described in the embodiment of the first aspect is performed.
Additional aspects and advantages of the application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the application.
Drawings
The application is further described with reference to the accompanying drawings and examples, in which:
FIG. 1 is a flow chart of a method for medical image segmentation in combination with anatomical priors according to one embodiment of the present application;
fig. 2 is a flowchart of step S130 in fig. 1;
FIG. 3 is a flowchart of step S220 in FIG. 2 in one embodiment;
FIG. 4 is a flow chart of step S320 in FIG. 3 in one embodiment;
FIG. 5 is a flowchart of step S330 in FIG. 3 in one embodiment;
fig. 6 is a flowchart of step S120 in fig. 1;
fig. 7 is a flowchart of step S620 in fig. 6;
Detailed Description
Embodiments of the present application are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals refer to the same or similar elements or elements having the same or similar functions throughout. The embodiments described below by referring to the drawings are exemplary only for the purpose of explaining the present application and are not to be construed as limiting the present application.
In the description of the present application, the description of the first and second is only for the purpose of distinguishing technical features, and should not be construed as indicating or implying relative importance or implying the number of technical features indicated or the precedence of the technical features indicated.
In the description of the present application, unless explicitly defined otherwise, terms such as arrangement, installation, connection, etc. should be construed broadly and the specific meaning of the terms in the present application can be reasonably determined by a person skilled in the art in combination with the specific contents of the technical solution.
In the description of the present application, a description with reference to the terms "one embodiment," "some embodiments," "illustrative embodiments," "examples," "specific examples," or "some examples," etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the present application. In this specification, schematic representations of the above terms do not necessarily refer to the same embodiments or examples. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
The embodiment of the application provides a medical image segmentation method combined with anatomical prior, which comprises the following steps: obtaining segmentation training data; the segmentation training data comprises medical image data and segmentation label data; obtaining priori information of the anatomical structure according to the segmentation training data; calculating an overall loss function according to the medical image data, the segmentation label data and the anatomical structure priori information; and training the medical image data according to the total loss function and a preset image segmentation model to obtain target image data.
As shown in fig. 1, fig. 1 is a flowchart of a medical image segmentation method combined with anatomical priors provided in some embodiments, and the medical image segmentation method combined with anatomical priors includes, but is not limited to, steps S110 to S140, specifically including:
s110, obtaining segmentation training data;
s120, obtaining priori information of the anatomical structure according to the segmentation training data;
s130, calculating an overall loss function according to medical image data, segmentation label data and anatomical structure priori information;
and S140, training the medical image data according to the total loss function and a preset image segmentation model to obtain target image data.
In step S110, the segmentation training data includes, but is not limited to, medical image data, segmentation label data.
In steps S120 to S140, the anatomical structure priori information includes, but is not limited to, anatomical structure priori knowledge of tissue organs, the human tissue organs have consistent anatomical structures, and clinicians often identify tissue structures in the medical image based on their anatomical structure priori knowledge of tissue organs, and this application proposes a solution: learning priori knowledge of anatomical structures of tissue organs to obtain priori information of the anatomical structures, guiding model training by using the obtained priori information of the anatomical structures, and finally segmenting target tissue organs in medical images by using a trained image segmentation model to obtain target image data.
According to the medical image segmentation method combining the anatomical priori, anatomical structure priori information is obtained according to the segmentation training data, anatomical structure characteristics of tissue and organs are used as the priori information, the overall loss function is calculated, the anatomical structure priori information is utilized to guide the training and optimization of a medical image segmentation model, further, target image data after image segmentation are obtained, and accuracy of influencing a segmentation result is improved.
According to some embodiments of the present application, computing an overall loss function from medical image data, segmentation label data, anatomical structure prior information includes: obtaining segmentation result data according to medical image data and a preset image segmentation model; and obtaining an overall loss function according to the segmentation label data, the anatomical structure priori information and the segmentation result data.
Fig. 2 is a flowchart of step S130 in some embodiments, step S130 illustrated in fig. 2 including, but not limited to, steps S210 to S220:
s210, obtaining segmentation result data according to medical image data and a preset image segmentation model;
s220, obtaining an overall loss function according to the segmentation label data, the anatomical structure priori information and the segmentation result data.
In a specific embodiment, after learning to obtain the priori anatomical structure information by using the self-encoder, the priori anatomical structure information is applied to training of an image segmentation model, wherein the image segmentation model is a depth network structure for image segmentation, and any depth network structure for image segmentation can be adopted.
In step S210, the medical image data including but not limited to a medical image is input into a preset image segmentation model to obtain corresponding segmentation result data. Wherein, the medical image data is marked as x i E, X, the image segmentation model is marked as S, and the segmentation result data is marked as m i s ＝S(x i ) I is the data index.
The segmentation result data is obtained through the calculation in the step S210 and is used in the calculation process of the total loss function in the step S220, so that the construction of an image segmentation model is realized, and the subsequent generation of medical image data is facilitated.
According to some embodiments of the present application, deriving an overall loss function from segmentation label data, anatomical structure prior information, segmentation result data, comprises: calculating segmentation loss information according to the segmentation label data and the segmentation result data; calculating priori loss information according to the anatomical structure priori information and the segmentation result data; and obtaining an overall loss function according to the segmentation loss information and the prior loss information.
Fig. 3 is a flow chart of step S220 in some embodiments, step S220 illustrated in fig. 3 including, but not limited to, steps S320 to S330:
s310, calculating segmentation loss information according to the segmentation label data and the segmentation result data;
s320, calculating priori loss information according to the anatomical structure priori information and the segmentation result data;
s330, obtaining an overall loss function according to the segmentation loss information and the prior loss information.
In step S310, division loss information corresponding to the division result data is calculated from the division tag data and the division tag data. Wherein the segmentation result data is recorded as m i s The method comprises the steps of carrying out a first treatment on the surface of the Dividing the tag data into m i g ∈M g The method comprises the steps of carrying out a first treatment on the surface of the The segmentation loss information is denoted as L seg 。
In step S320, the anatomical structure prior information is used to calculate prior loss information of the segmentation result according to the anatomical structure prior information of the segmentation target learned from the segmentation label data by the encoder E and the arbiter D. Wherein the segmentation result data is recorded as m i s The method comprises the steps of carrying out a first treatment on the surface of the The prior loss information is denoted as L pri 。
In step S330, the segmentation loss information is denoted as L seg The prior loss information is marked as L pri Obtaining the overall loss function L total 。
According to some embodiments of the present application, calculating prior loss information from the anatomical prior information, segmentation result data, includes: obtaining first sub priori loss information according to the segmentation result data and the segmentation label data; obtaining second sub priori loss information according to the segmentation result data; and obtaining priori loss information according to the first sub priori loss information and the second sub priori loss information.
Fig. 4 is a flow chart of step S320 in some embodiments, step S320 illustrated in fig. 4 including, but not limited to, steps S410 to S430:
s410, obtaining first sub priori loss information according to the segmentation result data and the segmentation label data;
s420, obtaining second sub priori loss information according to the segmentation result data;
and S430, obtaining priori loss information according to the first sub priori loss information and the second sub priori loss information.
In a particular embodiment, the information L is lost a priori pri The method consists of two parts, namely first sub priori loss information and second sub priori loss information.
The first sub-priori loss information is obtained in step S410, specifically: by calculating the segmentation result data m i s Is a coded vector E (m i s ) And all of the split tag data M g Is the encoded mean of (2)The distance between them is obtained, wherein I represents the amount of data; the second sub-a priori loss information is obtained in step S420, specifically: the overstepping device D divides the result data m i s Cross entropy of the classification output of (c).
According to some embodiments of the present application, deriving an overall loss function from the segmentation loss information, the a priori loss information, includes: acquiring loss weight information; and obtaining an overall loss function according to the segmentation loss information, the priori loss information and the loss weight information.
Fig. 5 is a flowchart of step S330 in some embodiments, step S330 illustrated in fig. 5 including, but not limited to, steps S510 to S520:
s510, acquiring loss weight information;
s520, obtaining an overall loss function according to the segmentation loss information, the priori loss information and the loss weight information.
In a specific embodiment, the overall loss function used to train the image segmentation model is defined as follows: overall loss function L total ＝L seg +λL pri . Wherein the segmentation loss information is denoted as L seg The prior loss information is denoted as L pri The overall loss function is denoted as L total λ is loss weight information. The anatomical structure priori information of the tissue and organ is introduced into the training process of the segmentation model through the total loss function, thereby realizing the medical image number by utilizing the anatomical structure priori informationThe segmentation is performed.
According to some embodiments of the present application, deriving anatomical structure prior information from segmentation training data includes: obtaining segmentation tag reconstruction data according to the segmentation tag data; and reconstructing data according to the segmentation labels to obtain prior information of the anatomical structure.
Fig. 6 is a flowchart of step S120 in some embodiments, step S120 illustrated in fig. 6 including, but not limited to, steps S610 through S620:
s610, obtaining segmentation tag reconstruction data according to the segmentation tag data;
s620, reconstructing data according to the segmentation labels to obtain prior information of the anatomical structure.
In a particular embodiment, the split tag data is denoted as M g The method comprises the steps of carrying out a first treatment on the surface of the The reconstruction data of the segmentation tag is the reconstruction result of the segmentation tag and is marked as M r The method comprises the steps of carrying out a first treatment on the surface of the Will divide the tag data M g The split tag reconstruction data M is obtained by inputting the split tag reconstruction data M into a self-encoder composed of an encoder E and a decoder R r ＝R(E(M g ) And then reconstructing data according to the segmentation tags to obtain prior information of the anatomical structure.
According to some embodiments of the present application, reconstructing data from segmentation labels yields anatomical prior information, including: calculating a coding loss function according to the segmentation tag data and the segmentation tag reconstruction data; and obtaining the priori information of the anatomical structure according to the coding loss function.
Fig. 7 is a flow chart of step S620 in some embodiments, and step S620 illustrated in fig. 7 includes, but is not limited to, steps S710 to S720:
s710, calculating a coding loss function according to the split tag data and the split tag reconstruction data;
s720, obtaining the priori information of the anatomical structure according to the coding loss function.
In a specific embodiment, the split tag data is input to a self-encoder consisting of an encoder and a decoder, and the split training data is denoted as (X, M) g ) Wherein X is medical image data, the label data M is to be segmented g And splitting the tag reconstruction data M r Input arbiter D, use D to M g And M r Distinguishing and calculating to obtain coding loss function L code Wherein the code loss function is a loss function of the self-encoder. According to the method, the countermeasure training of the self-encoder can be realized, the self-encoder is guided to optimize, learning of priori knowledge of the anatomical structure is realized, and the priori information of the anatomical structure is obtained according to the coding loss function.
The embodiment of the application provides a medical image segmentation device combined with anatomical prior, which comprises the following components: the segmentation training data acquisition module is used for acquiring segmentation training data; the segmentation training data comprises medical image data and segmentation label data; the priori information acquisition module is used for acquiring anatomical structure priori information according to the segmentation training data; the overall loss function acquisition module is used for calculating an overall loss function according to the medical image data, the segmentation tag data and the priori anatomical structure information; the target image data generation module is used for training the medical image data according to the total loss function and a preset image segmentation model to obtain target image data.
According to the medical image segmentation device combining the anatomical priori, the medical image segmentation method combining the anatomical priori is achieved, anatomical structure priori information is obtained according to segmentation training data, anatomical structure characteristics of tissue and organs are used as priori information, the overall loss function is calculated, the anatomical structure priori information is utilized to guide training and optimization of a medical image segmentation model, further target image data after image segmentation is obtained, and accuracy of influencing segmentation results is improved.
The embodiment of the application provides a medical image segmentation device combined with anatomical prior, which comprises the following components: memory, processor and computer program stored on the memory and executable on the processor, the processor implementing when executing the program: the medical image segmentation method according to any of the embodiments described above is applied in combination with anatomical prior.
According to an embodiment of the present application, a storage medium stores computer-executable instructions for: the medical image segmentation method in combination with anatomical prior of any of the embodiments described above is performed.
The above described apparatus embodiments are merely illustrative, wherein the units illustrated as separate components may or may not be physically separate, i.e. may be located in one place, or may be distributed over a plurality of network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
Those of ordinary skill in the art will appreciate that all or some of the steps, apparatus, and methods disclosed above may be implemented as software, firmware, hardware, and suitable combinations thereof. Some or all of the physical components may be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application specific integrated circuit. Such software may be distributed on computer readable media, which may include computer storage media (or non-transitory media) and communication media (or transitory media). The term computer storage media includes both volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data, as known to those skilled in the art. Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital Versatile Disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium which can be used to store the desired information and which can be accessed by a computer. Furthermore, as is well known to those of ordinary skill in the art, communication media typically embodies computer readable instructions, data structures, program modules or other data in a modulated data signal such as a carrier wave or other transport mechanism and includes any information delivery media.
The embodiments of the present application have been described in detail above with reference to the accompanying drawings, but the present application is not limited to the above embodiments, and various changes can be made within the knowledge of one of ordinary skill in the art without departing from the spirit of the present application. Furthermore, embodiments of the present application and features of the embodiments may be combined with each other without conflict.
Claims (8)
1. The medical image segmentation method combined with the anatomical prior is characterized by comprising the following steps of:
obtaining segmentation training data; the segmentation training data comprises medical image data and segmentation label data;
inputting the split tag data into a self-encoder to obtain split tag reconstruction data;
inputting the split tag data and the split tag reconstruction data into a discriminator, and calculating a coding loss function;
obtaining priori information of the anatomical structure according to the coding loss function;
calculating an overall loss function according to the medical image data, the segmentation label data and the anatomical structure prior information;
and training the medical image data according to the total loss function and a preset image segmentation model to obtain target image data.
2. The anatomical prior-combined medical image segmentation method according to claim 1, wherein the calculating an overall loss function from the medical image data, the segmentation label data, the anatomical prior information comprises:
obtaining segmentation result data according to the medical image data and the preset image segmentation model;
and obtaining the overall loss function according to the segmentation label data, the anatomical structure priori information and the segmentation result data.
3. The method of claim 2, wherein the obtaining the overall loss function according to the segmentation label data, the anatomical structure prior information, and the segmentation result data comprises:
calculating segmentation loss information according to the segmentation label data and the segmentation result data;
calculating priori loss information according to the anatomical structure priori information and the segmentation result data;
and obtaining the overall loss function according to the segmentation loss information and the prior loss information.
4. A medical image segmentation method in combination with anatomical prior according to claim 3, wherein the calculating prior loss information according to the anatomical prior information and the segmentation result data comprises:
obtaining first sub priori loss information according to the segmentation result data and the segmentation label data;
obtaining second sub priori loss information according to the segmentation result data;
and obtaining the priori loss information according to the first sub priori loss information and the second sub priori loss information.
5. A medical image segmentation method in combination with anatomical prior according to claim 3, wherein the deriving the overall loss function from the segmentation loss information and the prior loss information comprises:
acquiring loss weight information;
and obtaining the overall loss function according to the segmentation loss information, the priori loss information and the loss weight information.
6. The medical image segmentation apparatus is characterized by comprising:
the segmentation training data acquisition module is used for acquiring segmentation training data; the segmentation training data comprises medical image data and segmentation label data;
the prior information acquisition module is used for inputting the segmentation tag data into the self-encoder to obtain segmentation tag reconstruction data; inputting the split tag data and the split tag reconstruction data into a discriminator, and calculating a coding loss function; obtaining priori information of the anatomical structure according to the coding loss function;
the overall loss function acquisition module is used for calculating an overall loss function according to the medical image data, the segmentation tag data and the anatomical structure priori information;
and the target image data generation module is used for training the medical image data according to the overall loss function and a preset image segmentation model to obtain target image data.
7. The medical image segmentation apparatus is characterized by comprising: a memory, a processor, and a computer program stored on the memory and executable on the processor, the processor implementing when executing the program:
a medical image segmentation method in combination with anatomical priors as claimed in any one of claims 1 to 5.
8. A storage medium storing computer-executable instructions for:
a medical image segmentation method in combination with anatomical priors as claimed in any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110842739.7A CN113658187B (en) | 2021-07-26 | 2021-07-26 | Medical image segmentation method, device and storage medium combined with anatomical priori |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110842739.7A CN113658187B (en) | 2021-07-26 | 2021-07-26 | Medical image segmentation method, device and storage medium combined with anatomical priori |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113658187A CN113658187A (en) | 2021-11-16 |
| CN113658187B true CN113658187B (en) | 2024-03-29 |
Family
ID=78490114
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110842739.7A Active CN113658187B (en) | 2021-07-26 | 2021-07-26 | Medical image segmentation method, device and storage medium combined with anatomical priori |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113658187B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114693830B (en) * | 2022-05-27 | 2022-11-15 | 阿里巴巴达摩院(杭州)科技有限公司 | Multi-organ segmentation and model training method, equipment and medium for medical image |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110599491A (en) * | 2019-09-04 | 2019-12-20 | 腾讯医疗健康(深圳)有限公司 | Priori information-based eye image segmentation method, device, equipment and medium |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11132792B2 (en) * | 2018-02-22 | 2021-09-28 | Siemens Healthcare Gmbh | Cross domain medical image segmentation |
| US20210012486A1 (en) * | 2019-07-09 | 2021-01-14 | Shenzhen Malong Technologies Co., Ltd. | Image synthesis with generative adversarial network |
-
2021
- 2021-07-26 CN CN202110842739.7A patent/CN113658187B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110599491A (en) * | 2019-09-04 | 2019-12-20 | 腾讯医疗健康(深圳)有限公司 | Priori information-based eye image segmentation method, device, equipment and medium |
Non-Patent Citations (1)
| Title |
|---|
| 深度学习与解剖学先验融合的医学图像分割研究;郑寒;《中国优秀硕士学位论文全文数据库》(第第08期期);第i、ii、21-40、59-61页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113658187A (en) | 2021-11-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US9990712B2 (en) | Organ detection and segmentation | |
| JP6877868B2 (en) | Image processing equipment, image processing method and image processing program | |
| EP3108445B1 (en) | Sparse appearance learning-based segmentation | |
| CN108475428B (en) | System and method for heart model guided coronary artery segmentation | |
| US9305349B2 (en) | Apparatus and method for detecting lesion | |
| EP3664034B1 (en) | Method and data processing system for providing lymph node information | |
| US10878564B2 (en) | Systems and methods for processing 3D anatomical volumes based on localization of 2D slices thereof | |
| US11241190B2 (en) | Predicting response to therapy for adult and pediatric crohn's disease using radiomic features of mesenteric fat regions on baseline magnetic resonance enterography | |
| CN113159040B (en) | Method, device and system for generating medical image segmentation model | |
| US20200273167A1 (en) | Assessment of arterial calcifications | |
| CN112150472A (en) | Three-dimensional jaw bone image segmentation method and device based on CBCT (cone beam computed tomography) and terminal equipment | |
| CN113658187B (en) | Medical image segmentation method, device and storage medium combined with anatomical priori | |
| CN111128349A (en) | GAN-based medical image focus detection marking data enhancement method and device | |
| CN112950552B (en) | Rib segmentation marking method and system based on convolutional neural network | |
| CN111513743B (en) | Fracture detection method and device | |
| US20230222771A1 (en) | Method and system for automatic classification of radiographic images having different acquisition characteristics | |
| Kalapos et al. | Automated T1 and T2 mapping segmentation on cardiovascular magnetic resonance imaging using deep learning | |
| CN115294023A (en) | Liver tumor automatic segmentation method and device | |
| CN112862785B (en) | CTA image data identification method, device and storage medium | |
| Li et al. | Segmentation evaluation with sparse ground truth data: Simulating true segmentations as perfect/imperfect as those generated by humans | |
| CN114723710A (en) | Method and device for detecting ultrasonic video key frame based on neural network | |
| CN112862786B (en) | CTA image data processing method, device and storage medium | |
| CN113706541B (en) | Image processing method and device | |
| WO2023246937A1 (en) | Systems and methods for image processing | |
| CN112862787B (en) | CTA image data processing method, device and storage medium |
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 |