CN117198514A - Vulnerable plaque identification method and system based on CLIP model - Google Patents
Vulnerable plaque identification method and system based on CLIP model Download PDFInfo
- Publication number
- CN117198514A CN117198514A CN202311473361.3A CN202311473361A CN117198514A CN 117198514 A CN117198514 A CN 117198514A CN 202311473361 A CN202311473361 A CN 202311473361A CN 117198514 A CN117198514 A CN 117198514A
- Authority
- CN
- China
- Prior art keywords
- vulnerable plaque
- image
- model
- text
- clip
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 38
- PCHJSUWPFVWCPO-UHFFFAOYSA-N gold Chemical compound [Au] PCHJSUWPFVWCPO-UHFFFAOYSA-N 0.000 claims abstract description 5
- 230000036961 partial effect Effects 0.000 claims abstract description 5
- 230000008569 process Effects 0.000 claims abstract description 5
- 238000012545 processing Methods 0.000 claims description 25
- 230000003902 lesion Effects 0.000 claims description 12
- 238000007781 pre-processing Methods 0.000 claims description 12
- 238000012549 training Methods 0.000 claims description 10
- 238000004590 computer program Methods 0.000 claims description 9
- 208000031481 Pathologic Constriction Diseases 0.000 claims description 5
- 208000037804 stenosis Diseases 0.000 claims description 5
- 230000036262 stenosis Effects 0.000 claims description 5
- 238000000605 extraction Methods 0.000 claims description 4
- 230000006870 function Effects 0.000 claims description 4
- 238000010606 normalization Methods 0.000 claims description 4
- 238000005457 optimization Methods 0.000 claims description 4
- 238000012952 Resampling Methods 0.000 claims 1
- 230000002526 effect on cardiovascular system Effects 0.000 abstract description 6
- 238000011160 research Methods 0.000 abstract description 6
- 230000002411 adverse Effects 0.000 abstract description 5
- 230000002829 reductive effect Effects 0.000 abstract description 5
- 238000010801 machine learning Methods 0.000 description 4
- 235000009421 Myristica fragrans Nutrition 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000003628 erosive effect Effects 0.000 description 2
- 150000002632 lipids Chemical class 0.000 description 2
- 239000001115 mace Substances 0.000 description 2
- 239000003550 marker Substances 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 208000024891 symptom Diseases 0.000 description 2
- 208000004476 Acute Coronary Syndrome Diseases 0.000 description 1
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 1
- 244000270834 Myristica fragrans Species 0.000 description 1
- 208000007536 Thrombosis Diseases 0.000 description 1
- 230000003143 atherosclerotic effect Effects 0.000 description 1
- 230000002308 calcification Effects 0.000 description 1
- 230000007211 cardiovascular event Effects 0.000 description 1
- 238000002586 coronary angiography Methods 0.000 description 1
- 208000029078 coronary artery disease Diseases 0.000 description 1
- 210000004351 coronary vessel Anatomy 0.000 description 1
- 238000012937 correction 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
- 230000000873 masking effect Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000013508 migration Methods 0.000 description 1
- 230000005012 migration Effects 0.000 description 1
- 230000001338 necrotic effect Effects 0.000 description 1
- 230000001575 pathological effect Effects 0.000 description 1
- 238000004393 prognosis Methods 0.000 description 1
- 238000004445 quantitative analysis Methods 0.000 description 1
- 238000013139 quantization Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 230000002792 vascular Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Landscapes
- Image Analysis (AREA)
Abstract
The invention relates to the field of medical engineering, and provides a vulnerable plaque identification method and a vulnerable plaque identification system based on a CLIP model aiming at the influence of vulnerable plaque on main cardiovascular adverse events and the current situation of the current research of automatically identifying vulnerable plaque based on the image field. The vulnerable plaque identification network model constructed by the invention is based on the CLIP model, and a BN layer and a Dropout layer are introduced to respectively process text features and image features, so that the overfitting is reduced. In addition, considering that the characteristic judgment of partial vulnerable plaque is very subjective, the gold standard label is easy to mix with noise, bootstrapping loss is adopted to replace a standard cross entropy loss function, a predictive label is introduced into a bootstrapping loss formula, the loss value of a noise sample is reduced, the update of the noise sample is indirectly influenced, if the predictive value is a true value, the loss is 0, and the normal sample can still be effectively trained.
Description
Technical Field
The invention relates to the field of medical engineering, in particular to a vulnerable plaque identification method and a vulnerable plaque identification system based on a CLIP model.
Background
Patients with acute coronary syndromes have a high risk of death, and the main pathological basis is unstable plaque rupture or thrombosis secondary to plaque surface erosion. While unstable plaque rupture is closely related to plaque surface erosion (also known as vulnerable plaque or high risk plaque) and major cardiovascular adverse events (major adverse cardiovascular event, MACE). Thus, early identification of vulnerable plaque and, accordingly, enhanced intervention is of great importance in reducing MACE. Traditional invasive coronary angiography only can display vascular lumen conditions, and cannot directly display plaque and characteristics thereof. The coronary artery CTA (CCTA) can reliably evaluate the lumen stenosis and the functional significance thereof, can accurately evaluate the form and the composition of plaque, can identify vulnerable plaque, and has extremely important significance for guiding the clinical management of coronary heart disease patients.
In recent years, research into image histology and machine learning in the cardiovascular field has been increasing. The image histology reduces human measurement errors through an automatic plaque segmentation and quantization technology, can integrate clinical and image data simultaneously to comprehensively analyze diseases, and greatly improves the application value of a CCTA high-dimensional plaque quantitative analysis technology. For exploring the value of machine learning in coronary plaque, recent studies have shown that lesion features extracted with machine learning (including minimum luminal area, percent of atherosclerotic volume, fibrous fat and necrotic core volume, plaque volume, left anterior descending lesions, and reconstitution index) show higher prognostic value in MACEs within 5 years of prediction than the current predictive capabilities of cardiovascular risk score, coronary calcification score, and luminal stenosis severity, etc. Moreover, machine learning techniques can also be used for extraction of high-risk plaque features, identifying patients at risk of plaque progression. However, the current research of automatically identifying vulnerable plaques and guiding clinical treatment decisions and prognosis evaluation based on the image field is less, and the method is worthy of deep exploration and research, so that the method plays a greater value.
Disclosure of Invention
Aiming at the influence of vulnerable plaque on main cardiovascular adverse events and the current situation of the current research of automatically identifying vulnerable plaque based on the image field, the invention provides a vulnerable plaque identification method and system based on a CLIP model.
In order to achieve the above purpose, the present invention adopts the following technical scheme:
in a first aspect, the present invention provides a vulnerable plaque identification method based on a CLIP model, comprising the steps of:
s1, preprocessing an image;
s2, constructing and training a vulnerable plaque identification network model;
s3, inputting the preprocessed images and the preprocessed texts into a trained model, and completing identification and classification of vulnerable plaques.
Further, the specific process of image preprocessing in S1 is: CPR images with the known vessel center line and focus range are generated into CPR images with the spacing of 0.3 x 0.3, because vulnerable plaque symptoms are mainly lipid components, window width needs to be widened, window level 350HU is defined and used for normalization, window width 1500HU is used for normalization, focus box crop is taken from the normalized CPR images, and the size of 64 x 64 is resampled.
Further, the vulnerable plaque identification network model in the S2 comprises an image coding module, a text coding module, an image feature processing layer, a text feature processing layer, a feature splicing layer and a classifier; the image coding module is used for extracting picture features; the text coding module is used for extracting text characteristics; the image feature processing layer is used for processing the image features extracted by the image encoding module; the text feature processing layer is used for processing the text features extracted by the text encoding module; the feature splicing layer is used for splicing the processed image features and the text features; the classifier is used for outputting a recognition classification result.
Further, the image encoding module is a vision transformer or ResNet50 network; the image feature processing layer and the text feature processing layer are respectively a BN layer and a Dropout layer.
Further, the model training in the S2 adopts an AdamW parameter optimization method.
Further, the training is specifically:
considering that the characteristic judgment of partial vulnerable plaque is very subjective, the gold standard label is easy to mix with noise, and bootstrapping loss is adopted to replace a standard cross entropy loss function;
bootstrapping loss is shown below:
wherein:is a real label->Is a predictive tag,/->Is a predicted probability value,/>Is the weight of the noise and,Nis the number of samples.
Further, the text in S3 is a description of plaque in a general report, such as: plaque location, type, morphology, stenosis degree.
In a second aspect, the present invention provides a vulnerable plaque identification system based on a CLIP model, where the system is configured to implement the vulnerable plaque identification method based on the CLIP model described above, and the system includes an image preprocessing unit and a vulnerable plaque identification unit, where the image preprocessing unit is configured to generate CPR images with a spacing of 0.3×0.3 for MPR images with known vessel center lines and lesion ranges, normalize the CPR images with a window level of 350HU and a window width of 1500HU, take a lesion box crop for the normalized CPR images, and resample the lesion box crop to a size of 64×64; the vulnerable plaque identification unit comprises a vulnerable plaque identification network model, and mainly performs feature extraction, feature processing and feature splicing, so that vulnerable plaque identification classification results are obtained.
In a third aspect, the invention provides an electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing a vulnerable plaque identification method based on a CLIP model as described above when executing the computer program.
In a fourth aspect, the present invention provides a non-transitory computer readable storage medium having stored thereon a computer program for implementing a vulnerable plaque identification method based on a CLIP model as described above when executed by a processor.
Compared with the prior art, the invention has the following advantages:
aiming at the influence of vulnerable plaque on main cardiovascular adverse events and the current situation of the current research of automatically identifying vulnerable plaque based on the image field, the invention provides a vulnerable plaque identification method and system based on a CLIP model. The vulnerable plaque identification network model constructed by the invention is based on the CLIP model, and a BN layer and a Dropout layer are introduced to respectively process text features and image features, so that the overfitting is reduced. In addition, considering that the characteristic judgment of partial vulnerable plaque is very subjective, the gold standard label is easy to mix with noise, bootstrapping loss is adopted to replace a standard cross entropy loss function, a predictive label is introduced into a bootstrapping loss formula, the loss value of a noise sample is reduced, the update of the noise sample is indirectly influenced, if the predictive value is a true value, the loss is 0, and the normal sample can still be effectively trained.
Drawings
FIG. 1 is a flow chart of the method of the present invention.
Fig. 2 is a flowchart of image preprocessing.
FIG. 3 is a diagram of a vulnerable plaque identification network model of the present invention.
Fig. 4 is a schematic diagram of CNN architecture of a CLIP model image encoding module.
Fig. 5 is a schematic diagram of a CLIP model text encoder architecture.
Detailed Description
The technical scheme of the invention is specifically and specifically described below with reference to the embodiment of the invention and the attached drawings. It should be noted that the following examples are only for illustrating the present invention and are not intended to limit the scope of the present invention. The technical means used in the examples are conventional means well known to those skilled in the art unless otherwise indicated.
As shown in fig. 1, a vulnerable plaque identification method based on a CLIP model includes the following steps:
1. image preprocessing: CPR images with spacing of 0.3 x 0.3 are generated from MPR images with known vessel center lines and lesion ranges, because vulnerable plaque symptoms are mainly lipid components, a window width needs to be widened, the window width is defined as using window level 350HU and window width 1500HU for normalization, a patch is taken from a lesion box crop of the normalized CPR images, and the CPR images are resampled to 64 x 64 (the original aspect ratio is maintained, and the redundant place is supplemented with 0), as shown in figure 2.
2. Constructing and training a vulnerable plaque identification network model;
as shown in FIG. 3, the vulnerable plaque identification network model of the invention comprises an image coding module, a text coding module, an image feature processing layer, a text feature processing layer, a feature splicing layer and a classifier consisting of two full-connection layers; the image coding module is used for extracting picture features, the text coding module is used for extracting text features, the image feature processing layer is used for processing the image features extracted by the image coding module, the text feature processing layer is used for processing the text features extracted by the text coding module, the feature splicing layer splices the processed image features and the text features, and the classifier formed by the two fully-connected layers is used for outputting a recognition classification result.
The CLIP model is a visual model that trains a strong migration capability using a textual supervisory signal. The device consists of two coding modules which are respectively used for coding text data and image data. For the image coding module, a plurality of different model architectures are explored, two options of the traditional CNN architecture are shown in fig. 4, but training calculation efficiency of ViT variant of CLIP is 3 times higher, so ViT variant of CLIP is selected as the image encoder architecture. The text encoder is just a decoder-only transform, meaning that masking self-attention is used in each layer, and the Masked self-attention ensures that the representation of each marker in the sequence by the converter depends only on the marker preceding it. Fig. 5 is a basic description of a text encoder architecture.
The image coding module and the text coding module of the present invention are respectively an image coding module of a CLIP model (ViT variant of CLIP) and a text coding module. The image coding module may be a network such as vision transformer, resNet50, etc., and the text coding module inputs text as a description of the plaque in the report, possibly including plaque location, type, morphology, stenosis degree, etc.
In addition, in order to reduce the overfitting, the method introduces a BN layer and a Dropout layer to respectively process text features and image features, splice the obtained two features, and pass through a classifier of two full-connection layers.
The model training adopts an AdamW parameter optimization method, and the calculation efficiency of the AdamW parameter optimization method is higher than that of a traditional Adam optimizer.
The training is specifically as follows:
considering that the characteristic judgment of partial vulnerable plaque is very subjective, the gold standard label is easy to mix with noise, and bootstrapping loss is adopted to replace a standard cross entropy loss function;
cross entropy loss formula:
bootstrapping loss is shown below:
wherein:is a real label->Is a predictive tag,/->Is a predicted probability value,/>Is the weight of the noise and,Nis the number of samples.
The invention bootstrapping loss carries out loss correction on the basis of second-order cross entropy, and can be understood that the loss value calculated by a noise sample is larger, so that the influence of the noise sample on a model is larger, a predictive label is introduced into a bootstrapping loss formula, the loss value of the noise sample is reduced, the update of the influence is indirectly carried out, if the predictive value is a true value, the loss is 0, and the normal sample can still be effectively trained.
3. And inputting the preprocessed image and the text into a trained vulnerable plaque recognition network model to finish the recognition and classification of vulnerable plaque.
In another embodiment of the present invention, a vulnerable plaque identification system based on a CLIP model is provided, where the system is configured to implement the vulnerable plaque identification method based on the CLIP model described above, and the system includes an image preprocessing unit and a vulnerable plaque identification unit, where the image preprocessing unit is configured to generate CPR images with a spacing of 0.3×0.3 for MPR images with known vessel center lines and lesion ranges, normalize the CPR images with a window level of 350HU and a window width of 1500HU, take a lesion box crop for the normalized CPR images, and resample the lesion box crop to a size of 64×64; the vulnerable plaque identification unit comprises a vulnerable plaque identification network model, and performs feature extraction, feature processing and feature stitching, so that vulnerable plaque identification classification results are obtained.
In a third embodiment of the present invention, an electronic device is provided, including a memory, a processor, and a computer program stored on the memory and executable on the processor, the processor implementing a vulnerable plaque identification method based on a CLIP model as described above when executing the computer program.
A fourth embodiment of the present invention provides a non-transitory computer readable storage medium having stored thereon a computer program for implementing a vulnerable plaque identification method based on a CLIP model as described above when executed by a processor.
While the invention has been described in detail in the foregoing general description and with reference to specific embodiments thereof, it will be apparent to one skilled in the art that modifications and improvements can be made thereto. Accordingly, such modifications or improvements may be made without departing from the spirit of the invention and are intended to be within the scope of the invention as claimed.
Claims (10)
1. The vulnerable plaque identification method based on the CLIP model is characterized by comprising the following steps of:
s1, preprocessing an image;
s2, constructing and training a vulnerable plaque identification network model;
s3, inputting the preprocessed images and the preprocessed texts into a trained model, and completing identification and classification of vulnerable plaques.
2. The method for identifying vulnerable plaque based on CLIP model as claimed in claim 1, wherein the specific process of image preprocessing in S1 is: CPR images with known vessel center lines and lesion ranges are generated into CPR images with spacing of 0.3 x 0.3, window level 350HU and window width 1500HU are used for normalization, a patch is taken from the normalized CPR images, and the lesion box crop is resampled to 64 x 64 size.
3. The chip model-based vulnerable plaque identification method of claim 1, wherein the vulnerable plaque identification network model in S2 comprises an image coding module, a text coding module, an image feature processing layer, a text feature processing layer, a feature stitching layer and a classifier; the image coding module is used for extracting picture features, the text coding module is used for extracting text features, the image feature processing layer is used for processing the image features extracted by the image coding module, the text feature processing layer is used for processing the text features extracted by the text coding module, the feature splicing layer splices the processed image features with the text features, and the classifier is used for outputting recognition classification results.
4. The method for identifying vulnerable plaque based on CLIP model of claim 3 wherein the image coding module is a vision transformer or ResNet50 network; the image characteristic processing layer and the text characteristic processing layer are respectively a BN layer and a Dropout layer.
5. The method for identifying vulnerable plaque based on the CLIP model as claimed in claim 1, wherein the model S2 model training adopts AdamW parameter optimization method.
6. The method for identifying vulnerable plaque based on CLIP model of claim 5 wherein the training is specifically:
considering that the characteristic judgment of partial vulnerable plaque is very subjective, the gold standard label is easy to mix with noise, and bootstrapping loss is adopted to replace a standard cross entropy loss function;
bootstrapping loss is shown below:
；
wherein:is a real label->Is a predictive tag,/->Is a predicted probability value,/>Is the weight of the noise and,Nis the number of samples.
7. The method for identifying vulnerable plaque based on CLIP model as claimed in claim 1, wherein the text in S3 is a description of plaque in a general report, such as: plaque location, type, morphology, stenosis degree.
8. A vulnerable plaque identification system based on a CLIP model, which is characterized in that the system is used for realizing the vulnerable plaque identification method based on the CLIP model as claimed in any one of claims 1-7, the system comprises an image preprocessing unit and a vulnerable plaque identification unit, wherein the image preprocessing unit is used for generating a CPR image with spacing of 0.3 x 0.3 from an MPR image with a known blood vessel center line and focus range, normalizing the CPR image with window level 350HU and window width 1500HU, taking focus box crop of the normalized CPR image, and resampling the focus box crop to 64 x 64 size; the vulnerable plaque identification unit comprises a vulnerable plaque identification network model, and mainly performs feature extraction, feature processing and feature splicing, so that vulnerable plaque identification classification results are obtained.
9. An electronic device, characterized in that: comprising a memory, a processor and a computer program stored on the memory and executable on the processor, said processor implementing the vulnerable plaque identification method based on the CLIP model as claimed in any one of claims 1-7 when said computer program is executed.
10. A non-transitory computer readable storage medium characterized by: the medium has stored thereon a computer program for implementing a vulnerable plaque identification method based on a CLIP model as claimed in any of claims 1-7 when executed by a processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311473361.3A CN117198514B (en) | 2023-11-08 | 2023-11-08 | Vulnerable plaque identification method and system based on CLIP model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311473361.3A CN117198514B (en) | 2023-11-08 | 2023-11-08 | Vulnerable plaque identification method and system based on CLIP model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117198514A true CN117198514A (en) | 2023-12-08 |
| CN117198514B CN117198514B (en) | 2024-01-30 |
Family
ID=88989102
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311473361.3A Active CN117198514B (en) | 2023-11-08 | 2023-11-08 | Vulnerable plaque identification method and system based on CLIP model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117198514B (en) |
Citations (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103917166A (en) * | 2011-08-17 | 2014-07-09 | Vp诊断公司 | A method and system of characterization of carotid plaque |
| CN106056126A (en) * | 2015-02-13 | 2016-10-26 | 西门子公司 | Plaque vulnerability assessment in medical imaging |
| CN108492272A (en) * | 2018-03-26 | 2018-09-04 | 西安交通大学 | Cardiovascular vulnerable plaque recognition methods based on attention model and multitask neural network and system |
| US20190172197A1 (en) * | 2015-08-14 | 2019-06-06 | Elucid Bioimaging Inc. | Methods and systems for utilizing quantitative imaging |
| WO2019238976A1 (en) * | 2018-06-15 | 2019-12-19 | Université de Liège | Image classification using neural networks |
| US20200013163A1 (en) * | 2018-07-03 | 2020-01-09 | Caroline Choi | Method for Diagnosing, Predicting, Determining Prognosis, Monitoring, or Staging Disease Based on Vascularization Patterns |
| WO2021257893A1 (en) * | 2020-06-19 | 2021-12-23 | Cleerly, Inc. | Systems, methods, and devices for medical image analysis, diagnosis, risk stratification, decision making and/or disease tracking |
| CN114387464A (en) * | 2021-12-01 | 2022-04-22 | 杭州脉流科技有限公司 | Vulnerable plaque identification method based on IVUS image, computer device, readable storage medium and program product |
| US20220391755A1 (en) * | 2021-05-26 | 2022-12-08 | Salesforce.Com, Inc. | Systems and methods for vision-and-language representation learning |
| WO2023004451A1 (en) * | 2021-07-28 | 2023-02-02 | Artrya Limited | A coronary artery disease analysis system |
| CN115688028A (en) * | 2023-01-05 | 2023-02-03 | 杭州华得森生物技术有限公司 | Tumor cell growth state detection equipment |
| WO2023029817A1 (en) * | 2021-08-31 | 2023-03-09 | 北京字节跳动网络技术有限公司 | Medical report generation method and apparatus, model training method and apparatus, and device |
| CN116524593A (en) * | 2023-04-23 | 2023-08-01 | 北京建筑大学 | Dynamic gesture recognition method, system, equipment and medium |
| US20230255581A1 (en) * | 2020-09-02 | 2023-08-17 | The General Hospital Corporation | System for and method of deep learning diagnosis of plaque erosion through optical coherence tomography |
| CN116796251A (en) * | 2023-08-25 | 2023-09-22 | 江苏省互联网行业管理服务中心 | Poor website classification method, system and equipment based on image-text multi-mode |
-
2023
- 2023-11-08 CN CN202311473361.3A patent/CN117198514B/en active Active
Patent Citations (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103917166A (en) * | 2011-08-17 | 2014-07-09 | Vp诊断公司 | A method and system of characterization of carotid plaque |
| CN106056126A (en) * | 2015-02-13 | 2016-10-26 | 西门子公司 | Plaque vulnerability assessment in medical imaging |
| US20190172197A1 (en) * | 2015-08-14 | 2019-06-06 | Elucid Bioimaging Inc. | Methods and systems for utilizing quantitative imaging |
| CN108492272A (en) * | 2018-03-26 | 2018-09-04 | 西安交通大学 | Cardiovascular vulnerable plaque recognition methods based on attention model and multitask neural network and system |
| WO2019238976A1 (en) * | 2018-06-15 | 2019-12-19 | Université de Liège | Image classification using neural networks |
| US20200013163A1 (en) * | 2018-07-03 | 2020-01-09 | Caroline Choi | Method for Diagnosing, Predicting, Determining Prognosis, Monitoring, or Staging Disease Based on Vascularization Patterns |
| WO2021257893A1 (en) * | 2020-06-19 | 2021-12-23 | Cleerly, Inc. | Systems, methods, and devices for medical image analysis, diagnosis, risk stratification, decision making and/or disease tracking |
| US20230255581A1 (en) * | 2020-09-02 | 2023-08-17 | The General Hospital Corporation | System for and method of deep learning diagnosis of plaque erosion through optical coherence tomography |
| US20220391755A1 (en) * | 2021-05-26 | 2022-12-08 | Salesforce.Com, Inc. | Systems and methods for vision-and-language representation learning |
| WO2023004451A1 (en) * | 2021-07-28 | 2023-02-02 | Artrya Limited | A coronary artery disease analysis system |
| WO2023029817A1 (en) * | 2021-08-31 | 2023-03-09 | 北京字节跳动网络技术有限公司 | Medical report generation method and apparatus, model training method and apparatus, and device |
| CN114387464A (en) * | 2021-12-01 | 2022-04-22 | 杭州脉流科技有限公司 | Vulnerable plaque identification method based on IVUS image, computer device, readable storage medium and program product |
| CN115688028A (en) * | 2023-01-05 | 2023-02-03 | 杭州华得森生物技术有限公司 | Tumor cell growth state detection equipment |
| CN116524593A (en) * | 2023-04-23 | 2023-08-01 | 北京建筑大学 | Dynamic gesture recognition method, system, equipment and medium |
| CN116796251A (en) * | 2023-08-25 | 2023-09-22 | 江苏省互联网行业管理服务中心 | Poor website classification method, system and equipment based on image-text multi-mode |
Non-Patent Citations (3)
| Title |
|---|
| PEIWEN SHI 等: "WEAKLY SUPERVISED VULNERABLE PLAQUES DETECTION BY IVOCT IMAGE", 2020 IEEE 17TH INTERNATIONAL SYMPOSIUM ON BIOMEDICAL IMAGING (ISBI) * |
| 万黎明: "基于深度学习的HRMR图像颅内动脉粥样硬化斑块分割与识别方法研究", 中国优秀硕士论文电子期刊网 * |
| 门婷婷 等: "基于人工智能的成像技术在冠心病诊断中的应用", 《中国实验诊断学》, vol. 27, no. 8 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117198514B (en) | 2024-01-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112927240B (en) | CT image segmentation method based on improved AU-Net network | |
| CN109272048B (en) | Pattern recognition method based on deep convolutional neural network | |
| CN111476796B (en) | Semi-supervised coronary artery segmentation system and segmentation method combining multiple networks | |
| CN112446891B (en) | Medical image segmentation method based on U-Net network brain glioma | |
| CN111598875A (en) | Method, system and device for building thyroid nodule automatic detection model | |
| CN110751636B (en) | Fundus image retinal arteriosclerosis detection method based on improved coding and decoding network | |
| CN112381846A (en) | Ultrasonic thyroid nodule segmentation method based on asymmetric network | |
| CN112132800A (en) | Pulmonary fibrosis detection and severity evaluation method and system based on deep learning | |
| KR20200139606A (en) | Cervical cancer diagnosis system | |
| CN113610118A (en) | Fundus image classification method, device, equipment and medium based on multitask course learning | |
| CN110992309B (en) | Fundus image segmentation method based on deep information transfer network | |
| CN117036288A (en) | Tumor subtype diagnosis method for full-slice pathological image | |
| CN117351307B (en) | Model training method, device, equipment and storage medium | |
| Tan et al. | Lightweight pyramid network with spatial attention mechanism for accurate retinal vessel segmentation | |
| US6941288B2 (en) | Online learning method in a decision system | |
| Kumar et al. | Diabetic Retinopathy Detection using Deep Learning Methodology | |
| CN117198514B (en) | Vulnerable plaque identification method and system based on CLIP model | |
| CN113129310A (en) | Medical image segmentation system based on attention routing | |
| CN116740041A (en) | CTA scanning image analysis system and method based on machine vision | |
| CN116228731A (en) | Multi-contrast learning coronary artery high-risk plaque detection method, system and terminal | |
| Perumal et al. | Microaneurysms detection in fundus images using local fourier transform and neighbourhood analysis | |
| Ma et al. | [Retracted] Application of Artificial Intelligence in Cardiovascular Imaging | |
| CN115272357A (en) | Liver tumor segmentation method based on multi-temporal fusion and double-attention mechanism | |
| Wang et al. | PCRTAM-Net: A novel pre-activated convolution residual and triple attention mechanism network for retinal vessel segmentation | |
| UmaMaheswaran et al. | Enhanced non-contrast computed tomography images for early acute stroke detection using machine learning approach |
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 |