CN111415331A - Abnormality detection method and system based on category relation in positive chest radiograph - Google Patents
Abnormality detection method and system based on category relation in positive chest radiograph Download PDFInfo
- Publication number
- CN111415331A CN111415331A CN202010139192.XA CN202010139192A CN111415331A CN 111415331 A CN111415331 A CN 111415331A CN 202010139192 A CN202010139192 A CN 202010139192A CN 111415331 A CN111415331 A CN 111415331A
- Authority
- CN
- China
- Prior art keywords
- class
- category
- chest
- relation
- knowledge base
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
- G06F18/24133—Distances to prototypes
- G06F18/24137—Distances to cluster centroïds
- G06F18/2414—Smoothing the distance, e.g. radial basis function networks [RBFN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/29—Graphical models, e.g. Bayesian networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10081—Computed x-ray tomography [CT]
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Abstract
The invention discloses an abnormality detection method and system based on class relation in a positive chest film, which comprises the following steps: traversing all the positive chest film data, counting the co-occurrence frequency among all the anomalies, and establishing a co-occurrence frequency relation graph among the anomalies; representing the symbiotic relation of different types of exceptions by using the adjacency matrix, and generating a class knowledge base; mapping the category knowledge in the category knowledge base back to the prediction region by using the category probability vector to obtain the enhanced features of the prediction region; and splicing the enhanced features and the original features, and performing category classification and position regression. The invention applies the prior knowledge of class symbiosis to the detection process, and can obtain more accurate detection results.
Description
Technical Field
The invention relates to the technical field of medical image processing, in particular to an abnormality detection method and system based on class relation in a positive chest film.
Background
With the overall and rapid development of medical imaging technology, medical diagnosis modes are changed greatly, and more doctors select various medical images to make optional accurate diagnosis on the illness state of patients. In modern medicine, computer-aided medical image analysis techniques have been widely used. The diagnosis mode with the computer assistance can effectively reduce the workload of doctors and improve the working efficiency.
In recent years, with the rapid development of deep learning techniques, medical image analysis research based on convolutional neural networks is becoming mainstream. Chest x-ray scanning is a common type of chest disease examination, and a professional physician can identify and locate abnormal areas of the chest by means of chest images. Compared with the conventional machine vision method, the convolutional neural network has good feature extraction and expression capability, high detection precision and better interpretability.
In the thoracic region, different classes of abnormalities have some co-occurrence, and the presence of one abnormality often triggers another. Secondly, due to the complexity of human mechanisms, several factors can induce multiple abnormalities to occur simultaneously. Therefore, in a positive chest image, a plurality of abnormalities often exist at the same time, and the abnormalities are strongly related.
The prior chest radiography abnormity detection method rarely applies the prior knowledge to the detection process. The abnormality detection in the chest radiograph is to identify and position an abnormal region, and the prior knowledge of category symbiosis has good effectiveness in improving the detection precision of the abnormality.
Disclosure of Invention
In order to overcome the technical problems, the invention provides an abnormality detection method and system based on class relationship in a chest radiograph, which can obtain a more accurate detection result by applying the priori knowledge of class symbiosis to the detection process.
The technical scheme adopted for solving the technical problems is as follows:
a class relation-based abnormality detection method in a chest radiograph for orthostatic treatment comprises the following steps:
traversing all the positive chest film data, counting the co-occurrence frequency among all the anomalies, and establishing a co-occurrence frequency relation graph among the anomalies;
representing the symbiotic relation of different types of exceptions by using the adjacency matrix, and generating a class knowledge base;
mapping the category knowledge in the category knowledge base back to the prediction region by using the category probability vector to obtain the enhanced features of the prediction region;
and splicing the enhanced features and the original features, and performing category classification and position regression.
Optionally, counting the frequency of co-occurrence among all the anomalies according to the image marking information; and (3) obtaining the co-occurrence probability of each category and other categories by using a conditional probability function so as to generate a directed graph, wherein the graph relation is represented by an adjacency matrix.
Optionally, a global attention mechanism is introduced to perform sparse operation on the generated adjacency matrix; and sending the chest picture image into an attention network to obtain a category probability vector of the image.
Optionally, the feature extraction network in the convolutional layer and detection framework of the attention network shares parameters.
Optionally, after the feature map is obtained, performing a dimension reduction operation on the feature map by using a global average pooling operation; then, a fully-connected layer is connected behind the pooling layer to obtain a class probability vector.
Optionally, the class probability vector of each prediction region is obtained by using the weight parameter of the old classifier.
Optionally, the enhanced features and the original features are spliced to form a new feature expression, and the new feature expression is sent to a new classifier and a new regressor to perform class classification and position regression so as to obtain a better detection result.
The invention also provides an abnormality detection system based on the class relation in the orthostatic chest radiograph, which comprises the following steps:
the positive chest film data traversing module is used for traversing all the positive chest film data, counting the co-occurrence frequency among all the abnormalities and establishing a co-occurrence frequency relation graph among the abnormalities;
the class knowledge base generation module is used for representing the symbiotic relation of different classes of exceptions by utilizing the adjacency matrix so as to generate a class knowledge base;
the enhanced feature acquisition module is used for mapping the category knowledge in the category knowledge base back to the prediction region by using the probability vector to obtain the enhanced features of the prediction region;
and the classification and regression module is used for splicing the enhanced features and the original features to perform classification and position regression.
The invention also provides electronic equipment which comprises a memory, a processor and a computer program which is stored on the memory and can run on the processor, wherein the processor realizes the abnormality detection method based on the class relation in the orthopedics chest film when executing the program.
The invention also provides a computer readable storage medium, on which a computer program is stored, which when executed by a processor implements a method for detecting abnormalities in an orthotopic chest film based on class relationships.
The invention has the beneficial effects that:
1) the method constructs a category relation graph by counting the frequency of co-occurrence among the abnormalities in the chest radiography data; the category knowledge is shared and transmitted on the relational graph, and the prior knowledge is effectively added into the detection process by using the method, so that the detection process has better interpretability and strong applicability;
2) the invention provides an attention mechanism for eliminating the influence of information redundancy on the class knowledge transmission process; and sending the chest picture image into an attention network to obtain a global class probability vector so as to carry out sparse operation on the adjacent matrix. Attention is paid to the mechanism to highlight the learning key point, so that the detection accuracy is effectively improved;
3) the method maps the category knowledge in the category knowledge base to the corresponding prediction region through a mapping mechanism to form the enhanced feature of the region; the enhanced features are spliced with the original features and then sent to a new classifier and a regressor for classification and position regression, and the cascade operation of the new classifier and the old classifier and the regressor can effectively improve the robustness of the detection algorithm.
Drawings
FIG. 1 is a flow chart of the present invention;
FIG. 2 is a diagram illustrating the propagation of knowledge sharing by category according to the present invention.
Detailed Description
The first embodiment is as follows:
the abnormality detection method based on the category relationship in a chest radiograph according to the present invention will be described in detail below with reference to fig. 1 to 2 of the specification.
The embodiment provides an abnormality detection method based on a category relationship in an orthotopic chest radiograph, which comprises the following steps:
and step S1, traversing all the chest film data, counting the co-occurrence frequency among the abnormalities according to the image marking information, and establishing a co-occurrence frequency relation graph among the abnormalities.
Specifically, the co-occurrence frequency among all the category anomalies is counted, the co-occurrence probability of each category and other categories is obtained by using a conditional probability function, so that a directed graph is generated, and the graph relation can be represented by an adjacency matrix. Compared with an undirected graph, the directed graph is more reasonable in guiding the flow direction of the class knowledge and is more suitable for the condition that the frequency difference between two types of exceptions is large.
In step S2, a co-occurrence frequency map between anomalies is obtained in step S1. The symbiotic relationship of different types of exceptions can be represented by an adjacent matrix, and the class knowledge is shared and propagated by taking the adjacent matrix as a medium, so that a class knowledge base is generated finally.
When reasoning is carried out on a single chest radiography, the problem of information redundancy exists in an adjacent matrix based on all chest radiography data statistics. Therefore, a global attention mechanism is introduced to perform sparse operations on the current adjacency matrix.
And sending the chest picture image into an attention network to obtain a category probability vector of the image. Wherein, the feature extraction network in the convolution layer and detection frame of the attention network shares parameters. And after the feature map is obtained, performing dimension reduction operation on the feature map by using global average pooling operation.
A fully connected layer is followed after the pooling layer to obtain class probability vectors. The attention network and the detection network are synchronously trained, and a binary cross entropy loss function of the attention network is defined as:
wherein i represents a category index number; y isiA target tag representing category i; p is a radical ofiRepresenting the prediction probability of the category i. After the class probability vector of the whole image is obtained, sparse operation can be carried out on the adjacent matrix to obtain a sparse adjacent matrix.
As shown in fig. 2, the specific process of category knowledge propagation is as follows:
representing category semantic knowledge by using the weight of the old classifier; the trained classifier weights contain sufficient high-level semantic information to be represented as class features. And the class knowledge is shared and propagated by taking the sparse adjacent matrix as a medium, and finally, a global class knowledge base is generated.
In step S3, the global category knowledge base is obtained in step S2. Each prediction region has its own class probability vector, and the class probability vectors of the prediction regions can be obtained by using the weight parameters of the old classifier. And mapping the class knowledge in the class knowledge base back to the prediction region by using the class probability vector to serve as an enhanced feature of the region.
And finally, splicing the enhanced features and the original features to form a new feature expression, and sending the new feature expression into a new classifier and a new regressor to perform class classification and position regression so as to obtain a better detection result.
The embodiment applies the priori knowledge of class symbiosis to the detection process, and can obtain a more accurate detection result.
Example two:
the embodiment provides an abnormality detection system based on a category relationship in an orthotopic chest radiograph, which comprises: the positive chest film data traversing module is used for traversing all the positive chest film data, counting the co-occurrence frequency among all the abnormalities and establishing a co-occurrence frequency relation graph among the abnormalities;
the class knowledge base generation module is used for representing the symbiotic relation of different classes of exceptions by utilizing the adjacency matrix so as to generate a class knowledge base;
the enhanced feature acquisition module is used for mapping the category knowledge in the category knowledge base back to the prediction region by using the probability vector to obtain the enhanced features of the prediction region;
and the classification and regression module is used for splicing the enhanced features and the original features to perform classification and position regression.
Example three:
the present embodiment provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor executes the program to implement the method for detecting abnormality in an orthotopic chest film based on the category relationship according to the first embodiment.
Example four:
the present embodiment provides a computer-readable storage medium, on which a computer program is stored, which when executed by a processor implements a method for detecting abnormality based on a class relationship in an orthotopic chest film according to the first embodiment.
The steps involved in the second to fourth embodiments correspond to the first embodiment of the method, and the detailed description thereof can be found in the relevant description of the first embodiment.
The term "computer-readable storage medium" should be taken to include a single medium or multiple media containing one or more sets of instructions; it should also be understood to include any medium that is capable of storing, encoding or carrying a set of instructions for execution by a processor and that cause the processor to perform any of the methods of the present invention.
The above description is only an embodiment of the present invention, and variations that can be easily conceived by those skilled in the art within the scope of the present invention are also included in the scope of the present invention.
Claims (10)
1. A method for detecting abnormality based on class relation in a positive chest film is characterized by comprising the following steps:
traversing all the positive chest film data, counting the co-occurrence frequency among all the anomalies, and establishing a co-occurrence frequency relation graph among the anomalies;
representing the symbiotic relation of different types of exceptions by using the adjacency matrix, and generating a class knowledge base;
mapping the category knowledge in the category knowledge base back to the prediction region by using the category probability vector to obtain the enhanced features of the prediction region;
and splicing the enhanced features and the original features, and performing category classification and position regression.
2. The method of claim 1, wherein the number of co-occurrences between abnormalities is counted according to image labeling information; and (3) obtaining the co-occurrence probability of each category and other categories by using a conditional probability function so as to generate a directed graph, wherein the graph relation is represented by an adjacency matrix.
3. The abnormality detection method based on the class relationship in the orthotopic chest radiograph as claimed in claim 1 or 2, characterized in that a global attention mechanism is introduced to perform sparse operation on the generated adjacency matrix; and sending the chest picture image into an attention network to obtain a category probability vector of the image.
4. The method of claim 3, wherein the convolutional layer of the attention network shares parameters with the feature extraction network in the detection framework.
5. The method for detecting abnormality in an orthotopic chest radiograph based on class relationship as claimed in claim 3, wherein after the feature map is obtained, a global average pooling operation is used to perform a dimension reduction operation on the feature map; then, a fully-connected layer is connected behind the pooling layer to obtain a class probability vector.
6. The method as claimed in claim 1, wherein the class probability vector of each prediction region is obtained by using the weight parameters of the old classifier.
7. The method for detecting abnormality in an orthotopic chest radiograph based on class relationship as claimed in claim 1, wherein the enhanced features and the original features are spliced to form a new feature expression, and the new feature expression is sent to a new classifier and a new regressor for class classification and position regression to obtain a better detection result.
8. A system for detecting abnormalities in an orthotopic chest film based on class relationships, comprising:
the positive chest film data traversing module is used for traversing all the positive chest film data, counting the co-occurrence frequency among all the abnormalities and establishing a co-occurrence frequency relation graph among the abnormalities;
the class knowledge base generation module is used for representing the symbiotic relation of different classes of exceptions by utilizing the adjacency matrix so as to generate a class knowledge base;
the enhanced feature acquisition module is used for mapping the category knowledge in the category knowledge base back to the prediction region by using the probability vector to obtain the enhanced features of the prediction region;
and the classification and regression module is used for splicing the enhanced features and the original features to perform classification and position regression.
9. 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 implement a method for detecting abnormalities in an orthotopic chest film based on class relationships as claimed in any one of claims 1 to 7.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, implements a method for detecting abnormalities in an orthotopic chest film based on class relationships as claimed in any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010139192.XA CN111415331B (en) | 2020-03-03 | 2020-03-03 | Abnormal detection method and system based on category relation in positive chest radiography |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010139192.XA CN111415331B (en) | 2020-03-03 | 2020-03-03 | Abnormal detection method and system based on category relation in positive chest radiography |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111415331A true CN111415331A (en) | 2020-07-14 |
| CN111415331B CN111415331B (en) | 2023-05-23 |
Family
ID=71492764
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010139192.XA Active CN111415331B (en) | 2020-03-03 | 2020-03-03 | Abnormal detection method and system based on category relation in positive chest radiography |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111415331B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112382360A (en) * | 2020-12-03 | 2021-02-19 | 卫宁健康科技集团股份有限公司 | Automatic generation system of diagnosis report, storage medium and electronic equipment |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090208071A1 (en) * | 2005-02-15 | 2009-08-20 | Olympus Corporation | Medical Image Processing Apparatus, Luminal Image Processing Apparatus, Luminal Image Processing Method, and Programs for the Same |
| CN101710932A (en) * | 2009-12-21 | 2010-05-19 | 深圳华为通信技术有限公司 | Image stitching method and device |
| CN105513057A (en) * | 2015-11-30 | 2016-04-20 | 首都医科大学 | Tumor auxiliary diagnosis method |
| CN106600654A (en) * | 2017-01-24 | 2017-04-26 | 浙江四点灵机器人股份有限公司 | Large viewing angle depth camera splicing device and splicing method |
| CN107730484A (en) * | 2017-07-26 | 2018-02-23 | 江西中科九峰智慧医疗科技有限公司 | A kind of abnormal rabat intelligent identification Method and system based on deep learning |
| CN108197109A (en) * | 2017-12-29 | 2018-06-22 | 北京百分点信息科技有限公司 | A kind of multilingual analysis method and device based on natural language processing |
| US20180205889A1 (en) * | 2017-01-13 | 2018-07-19 | Gopro, Inc. | Apparatus and methods for the storage of overlapping regions of imaging data for the generation of optimized stitched images |
| CN109408743A (en) * | 2018-08-21 | 2019-03-01 | 中国科学院自动化研究所 | Text link embedding grammar |
| CN110033019A (en) * | 2019-03-06 | 2019-07-19 | 腾讯科技（深圳）有限公司 | Method for detecting abnormality, device and the storage medium of human body |
| CN110111325A (en) * | 2019-05-14 | 2019-08-09 | 深圳大学 | Neuroimaging classification method, terminal and computer readable storage medium |
| CN110335267A (en) * | 2019-07-05 | 2019-10-15 | 华侨大学 | A kind of cervical lesions method for detecting area |
| CN110555825A (en) * | 2019-07-23 | 2019-12-10 | 北京赛迈特锐医疗科技有限公司 | Intelligent diagnostic system and diagnostic method for chest X-ray image |
| CN110689044A (en) * | 2019-08-22 | 2020-01-14 | 湖南四灵电子科技有限公司 | Target detection method and system combining relationship between targets |
-
2020
- 2020-03-03 CN CN202010139192.XA patent/CN111415331B/en active Active
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090208071A1 (en) * | 2005-02-15 | 2009-08-20 | Olympus Corporation | Medical Image Processing Apparatus, Luminal Image Processing Apparatus, Luminal Image Processing Method, and Programs for the Same |
| CN101710932A (en) * | 2009-12-21 | 2010-05-19 | 深圳华为通信技术有限公司 | Image stitching method and device |
| CN105513057A (en) * | 2015-11-30 | 2016-04-20 | 首都医科大学 | Tumor auxiliary diagnosis method |
| US20180205889A1 (en) * | 2017-01-13 | 2018-07-19 | Gopro, Inc. | Apparatus and methods for the storage of overlapping regions of imaging data for the generation of optimized stitched images |
| CN106600654A (en) * | 2017-01-24 | 2017-04-26 | 浙江四点灵机器人股份有限公司 | Large viewing angle depth camera splicing device and splicing method |
| CN107730484A (en) * | 2017-07-26 | 2018-02-23 | 江西中科九峰智慧医疗科技有限公司 | A kind of abnormal rabat intelligent identification Method and system based on deep learning |
| CN108197109A (en) * | 2017-12-29 | 2018-06-22 | 北京百分点信息科技有限公司 | A kind of multilingual analysis method and device based on natural language processing |
| CN109408743A (en) * | 2018-08-21 | 2019-03-01 | 中国科学院自动化研究所 | Text link embedding grammar |
| CN110033019A (en) * | 2019-03-06 | 2019-07-19 | 腾讯科技（深圳）有限公司 | Method for detecting abnormality, device and the storage medium of human body |
| CN110111325A (en) * | 2019-05-14 | 2019-08-09 | 深圳大学 | Neuroimaging classification method, terminal and computer readable storage medium |
| CN110335267A (en) * | 2019-07-05 | 2019-10-15 | 华侨大学 | A kind of cervical lesions method for detecting area |
| CN110555825A (en) * | 2019-07-23 | 2019-12-10 | 北京赛迈特锐医疗科技有限公司 | Intelligent diagnostic system and diagnostic method for chest X-ray image |
| CN110689044A (en) * | 2019-08-22 | 2020-01-14 | 湖南四灵电子科技有限公司 | Target detection method and system combining relationship between targets |
Non-Patent Citations (3)
| Title |
|---|
| 俞益洲等: "深度学习在医学影像分析中的应用综述" * |
| 李?;王艳;马宗庆;张波;罗红;周激流;: "基于DenseASPP模型的超声图像分割" * |
| 韩坤;潘海为;张伟;边晓菲;陈春伶;何舒宁;: "基于多模态医学图像的Alzheimer病分类方法" * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112382360A (en) * | 2020-12-03 | 2021-02-19 | 卫宁健康科技集团股份有限公司 | Automatic generation system of diagnosis report, storage medium and electronic equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111415331B (en) | 2023-05-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Mahmood et al. | Deep adversarial training for multi-organ nuclei segmentation in histopathology images | |
| Pesce et al. | Learning to detect chest radiographs containing pulmonary lesions using visual attention networks | |
| US10691980B1 (en) | Multi-task learning for chest X-ray abnormality classification | |
| Saha et al. | Retracted article: Graphcovidnet: A graph neural network based model for detecting COVID-19 from ct scans and x-rays of chest | |
| US9892361B2 (en) | Method and system for cross-domain synthesis of medical images using contextual deep network | |
| US10853449B1 (en) | Report formatting for automated or assisted analysis of medical imaging data and medical diagnosis | |
| Rajaraman et al. | Analyzing inter-reader variability affecting deep ensemble learning for COVID-19 detection in chest radiographs | |
| Azizi et al. | Robust and efficient medical imaging with self-supervision | |
| US10692602B1 (en) | Structuring free text medical reports with forced taxonomies | |
| Sirjani et al. | Automatic cardiac evaluations using a deep video object segmentation network | |
| CN113569891A (en) | Training data processing device, electronic equipment and storage medium of neural network model | |
| Shamrat et al. | Analysing most efficient deep learning model to detect COVID-19 from computer tomography images | |
| CN111415331B (en) | Abnormal detection method and system based on category relation in positive chest radiography | |
| Su et al. | Attention2Minority: A salient instance inference-based multiple instance learning for classifying small lesions in whole slide images | |
| CN114612484B (en) | Retina OCT image segmentation method based on unsupervised learning | |
| Zhou et al. | Lung swapping autoencoder: Learning a disentangled structure-texture representation of chest radiographs | |
| Hao et al. | MFUnetr: A transformer-based multi-task learning network for multi-organ segmentation from partially labeled datasets | |
| Chaisangmongkon et al. | External validation of deep learning algorithms for cardiothoracic ratio measurement | |
| CN116982038A (en) | Image construction and visualization of multiple immunofluorescence images | |
| Roth et al. | Mask mining for improved liver lesion segmentation | |
| Rifai et al. | Analysis for diagnosis of pneumonia symptoms using chest X-ray based on MobileNetV2 models with image enhancement using white balance and contrast limited adaptive histogram equalization (CLAHE) | |
| Mann et al. | A hybrid deep convolutional neural network model for improved diagnosis of pneumonia | |
| Pérez-García et al. | RadEdit: stress-testing biomedical vision models via diffusion image editing | |
| Chakravarthi et al. | Sparsely Supervised Learning for Medical Image Classification on Noisy Heterogeneous Data | |
| Subashini et al. | Detection of Covid-19 using Mask R-CNN |
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 |