CN108198625B - Deep learning method and device for analyzing high-dimensional medical data - Google Patents
Deep learning method and device for analyzing high-dimensional medical data Download PDFInfo
- Publication number
- CN108198625B CN108198625B CN201611122716.4A CN201611122716A CN108198625B CN 108198625 B CN108198625 B CN 108198625B CN 201611122716 A CN201611122716 A CN 201611122716A CN 108198625 B CN108198625 B CN 108198625B
- Authority
- CN
- China
- Prior art keywords
- data
- length
- feature extraction
- mode
- dimensional medical
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16Z—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS, NOT OTHERWISE PROVIDED FOR
- G16Z99/00—Subject matter not provided for in other main groups of this subclass
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/70—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for mining of medical data, e.g. analysing previous cases of other patients
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Public Health (AREA)
- Medical Informatics (AREA)
- Pathology (AREA)
- Databases & Information Systems (AREA)
- Epidemiology (AREA)
- Primary Health Care (AREA)
- Image Analysis (AREA)
- Measuring And Recording Apparatus For Diagnosis (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
The application discloses a deep learning method and device for analyzing high-dimensional medical data. The method comprises the following steps: reading high-dimensional medical data, and preprocessing the high-dimensional medical data into a data format which can be received by a feature extraction model by analyzing data attributes; selecting different feature extraction models according to different preprocessing modes, inputting preprocessed data into the selected feature extraction models for feature extraction, and obtaining corresponding feature vectors; and inputting the corresponding characteristic vector into a recurrent neural network model for training to obtain a final deep learning model for medical data analysis. The method can improve the accuracy of high-dimensional medical data analysis based on the deep learning model. The present invention also includes a deep learning apparatus for analyzing high-dimensional medical data, comprising: the device comprises a preprocessing module, a feature extraction module and a model training module.
Description
Technical Field
The invention relates to the field of medical artificial intelligence and big data processing, in particular to a method and a device for analyzing high-dimensional medical data.
Background
In recent years, the artificial intelligence technology is developed vigorously, and with the rise of a new artificial intelligence technology taking a deep learning framework as an inner core, the development and the promotion of the technology are greatly achieved in various fields, and technologies such as AlphaGo, unmanned vehicles and voice recognition, which are expected for many years, are broken through in a short time. In the visible future, the deep learning also promotes the development of big data analysis and artificial intelligence application in the medical industry, and the deep learning method has great potential in changing medical health. The deep learning method is an artificial intelligence method for multi-level feature learning by constructing a deep network structure, and is widely and effectively used in the fields of image recognition, voice recognition and the like. With the breakthrough progress of the deep learning method on large-scale image classification (ImageNet), the deep learning is greatly concerned in all aspects, and the deep learning is successful in the fields of image recognition and voice recognition.
Particularly in the medical industry, the dimensionality of the medical field is higher than that of a common application scene, diagnosis and treatment data of each patient are complex, a deep learning and data analysis model is huge and complex than that of the common model, the training cost is very high, a large amount of manpower and material resources are consumed for learning and training the deep learning, big data and machine learning model by using a traditional method, and the economic feasibility of the application is greatly reduced. The high-dimensional medical image relates to multi-dimensional data such as CT, PET, SPECT, MRI, fMRI and the like, even for a single patient, the data volume is very large, and doctors take time and labor to process the image data, so the deep learning-based analysis method can greatly reduce the workload of the doctors and assist the doctors to more effectively finish the diagnosis of the patients.
Disclosure of Invention
In order to overcome the defects of the prior art, the invention provides a deep learning method and a deep learning device for analyzing high-dimensional medical data, which can effectively solve the problem of analyzing and processing the high-dimensional medical data so as to improve the practical effect of deep learning in the field of processing the medical data.
The invention discloses a deep learning method for analyzing high-dimensional medical data, which comprises the following steps:
s1: reading high-dimensional medical data, and preprocessing the high-dimensional medical data into a data format which can be received by a feature extraction model by analyzing data attributes;
s2: selecting different feature extraction models according to different preprocessing modes, and performing feature extraction on preprocessed data through the selected feature extraction models to obtain corresponding feature vectors;
s3: training the corresponding feature vectors to obtain a final deep learning model for medical data analysis;
in step S1, the method of preprocessing the high-dimensional data includes: a fixed-length interval extraction mode, a fixed-length overlapping extraction mode and a variable-length sequence processing mode.
Preferably, the fixed-length interval extraction mode includes: and selecting data with different lengths according to a certain interval, and finally forming a data format with a certain length for a subsequent feature extraction model.
Preferably, the fixed-length overlapping extraction manner includes: for data with different lengths, a certain number of frame numbers are extracted from one end of the data to serve as a first group of data, then the same number of frame numbers are extracted to serve as a second group of data, the two groups of data are partially overlapped, and finally a data format with a certain length is formed for subsequent feature extraction models.
Preferably, the variable length sequence processing method includes: and for data with different lengths, adding the data into a predetermined data format with a certain length, and if the length of the data is smaller than the fixed length, adding 0 to the subsequent data for alignment so as to be used by a subsequent feature extraction model.
Preferably, in step S2, for the data format obtained by the fixed-length interval extraction method and the variable-length sequence processing method, the 2d convolutional neural network or the 2d cyclic neural network is selected, and for the data format obtained by the fixed-length overlap extraction method, the 3d convolutional neural network or the 3d cyclic neural network is selected.
The present invention also relates to a deep learning apparatus for analyzing high-dimensional medical data, which includes: the data preprocessing module reads the high-dimensional medical data, and preprocesses the high-dimensional medical data into a data format which can be received by the feature extraction model by analyzing the data attributes; the feature extraction module is used for receiving the preprocessed data and extracting features to obtain corresponding feature vectors, wherein the feature extraction model is selected according to different preprocessing modes; the model training module receives and trains the corresponding characteristic vectors to obtain a final deep learning model for medical data analysis; the method for preprocessing the high-dimensional data comprises the following steps: a fixed-length interval extraction mode, a fixed-length overlapping extraction mode and a variable-length sequence processing mode.
Preferably, for the data format obtained by the fixed-length interval extraction mode and the variable-length sequence processing mode, a 2d convolutional neural network or a 2d cyclic neural network is selected as the feature extraction module, and for the data format obtained by the fixed-length overlap extraction mode, a 3d convolutional neural network or a 3d cyclic neural network is selected as the feature extraction module.
The technical scheme provided by the invention has the following beneficial effects: the accuracy of high-dimensional medical data analysis based on the deep learning model can be improved, so that the method is used for analyzing medical images and has good application value in the aspect of intelligent diagnosis of the medical images.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. It is obvious that the drawings in the following description are only some embodiments of the invention, and that for a person skilled in the art, other drawings can be derived from them without inventive effort.
FIG. 1 is a schematic diagram of a deep learning method of analyzing high-dimensional medical data according to an embodiment of the invention;
FIG. 2 is a flow diagram of a deep learning method of analyzing high-dimensional medical data according to an embodiment of the invention;
fig. 3 is a block diagram of a deep learning apparatus for analyzing high-dimensional medical data according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, embodiments of the present invention will be described in detail with reference to the accompanying drawings.
The invention provides a deep learning method for analyzing high-dimensional medical data, which is a schematic diagram of the deep learning method for analyzing high-dimensional medical data according to an embodiment of the invention, as shown in fig. 1.
Firstly, reading high-dimensional medical data, and preprocessing the high-dimensional medical data into a data format which can be received by a feature extraction model by analyzing data attributes; the feature extraction model performs feature extraction on the preprocessed data to obtain feature vectors; and training by using the feature vectors to obtain a final deep learning model so as to be used for analyzing the medical data.
As shown in fig. 2, a deep learning method for analyzing high-dimensional medical data according to the present invention includes the following steps:
s1: and reading the high-dimensional medical data, and preprocessing the high-dimensional medical data into a data format which can be received by the feature extraction model by analyzing the data attribute.
S2: and selecting different feature extraction models according to different preprocessing modes, inputting the preprocessed data into the selected feature extraction model for feature extraction, and obtaining corresponding feature vectors.
S3: and inputting the corresponding characteristic vector into a recurrent neural network model for training to obtain a final deep learning model for medical data analysis.
In step S1, the method of preprocessing the high-dimensional data includes: a fixed-length interval extraction mode, a fixed-length overlapping extraction mode and a variable-length sequence processing mode.
The fixed-length interval extraction mode comprises the following steps: taking data with different lengths, such as CT slice data with different layers as a reference, taking the middle layer as a reference, respectively selecting slice data from two sides at a certain interval, and finally forming a data format (100, 1, 512, 512) with a specific length (for example, 100 slice layers) for subsequent feature extraction.
The fixed-length overlapping extraction mode comprises the following steps: for data with different lengths, a certain number of frames are extracted from one end of the data to serve as a first group of data, then the same number of frames are extracted to serve as a second group of data, the two groups of data are overlapped in a certain amount, and finally a data format (100, 50, 512, 512) with a specific length is formed for subsequent feature extraction.
The variable length sequence processing mode comprises the following steps: and for data with different lengths, adding all slice layers of the data into a predetermined data format with a certain length, and if the length of the data is smaller than the fixed length, adding 0 to the following data for alignment. The final data format is (500, 1, 512, 512) for subsequent feature extraction.
Selecting a feature extraction model according to different data formats, and performing feature extraction to obtain corresponding feature vectors, wherein for the data formats (100, 1, 512, 512) obtained by processing in a fixed-length interval extraction mode, for the data formats (500, 1, 512, 512) obtained by processing in a variable-length sequence processing mode, a 2d Convolutional Neural Network (CNN) or a 2d cyclic (recursive) neural network (RNN) and the like are selected to perform feature extraction to obtain the feature vectors, and the formats (100, 4096) or (500, 4096) are obtained. And (3) for the data format (100, 50, 512, 512) obtained by the fixed-length overlapping extraction mode, selecting a convolution neural network of 3d or a circulation (recursive) neural network of 3d to perform feature extraction to obtain a feature vector, wherein the format (100, 4096) is obtained.
As shown in fig. 2, the present invention also relates to a deep learning apparatus for analyzing high-dimensional medical data, the apparatus comprising:
and the data preprocessing module reads the high-dimensional medical data and preprocesses the high-dimensional medical data into a data format which can be received by the feature extraction model by analyzing the data attributes.
And the feature extraction module selects different feature extraction models according to different preprocessing modes, inputs the preprocessed data into the selected feature extraction model for feature extraction, and obtains corresponding feature vectors.
And the model training module inputs the corresponding characteristic vectors into the recurrent neural network model for training to obtain a final deep learning model for medical data analysis.
In the data preprocessing module, the preprocessing mode includes: a fixed-length interval extraction mode, a fixed-length overlapping extraction mode and a variable-length sequence processing mode.
In the feature extraction module, different feature extraction models are selected for feature extraction according to different preprocessing modes.
The following description is given by way of example.
The first embodiment: the deep learning device provided by the invention is used for deep learning the high-dimensional medical data of the lung CT image.
A preprocessing module: it reads the high-dimensional medical data information of lung CT image, and because the lung CT scanning interval is different, the number of all slices is different every CT. For the CT of (300, 1, 512, 512), the slice size is 512 × 512, the number of slices is 300, if 100 frames are to be extracted for analysis, the fixed-length interval extraction method is adopted to process: extracting 1 frame for analysis every 3 frames, namely, the fixed interval is 2 frames, thus obtaining 100 frames of image data of the features to be extracted;
a feature extraction module: for the fixed-length interval extraction mode, selecting a 2d feature extraction model, inputting the preprocessed (100, 1, 512, 512) into the 2d feature extraction model, and obtaining a (100, 4096) format feature vector;
a model training module: inputting the feature vectors (N, 100, 4096) with the number of the training sets N into a recurrent neural network for training to obtain a final deep learning model.
Second embodiment: the deep learning device provided by the invention is used for deep learning the high-dimensional medical data of the head MRI image.
A data processing module: which reads high-dimensional medical data information of the MRI image of the head; for example, for an MRI of (416, 1, 512, 512), the slice size is 512 × 512, the number of slices is 416, and the processing is performed by using a fixed-length overlap extraction method: extracting 20 frames from one end of the data to be used as a first group of data, moving the step length by 4, taking 20 frames to be used as a second group of data, namely overlapping 16 frames in each group of data, and processing to obtain (416-20)/4+1 which is 100 groups of data of the features to be extracted;
a feature extraction module: selecting a 3d feature extraction model for the data processed by the fixed-length overlapping extraction mode, inputting the preprocessed (100, 20, 512, 512) into the 3d feature extraction model, and obtaining a (100, 4096) format feature vector;
a model training module: inputting the feature vectors (N, 100, 4096) with the number of the training sets N into a recurrent neural network for training to obtain a final deep learning model.
The third embodiment: the deep learning device provided by the invention is used for deep learning the high-dimensional medical data of the lung CT image.
A data processing module: it reads the high-dimensional medical data information of lung CT image, and because the lung CT scanning interval is different, the number of all slices is different every CT. For CT of different lengths such as (300, 1, 512, 512), (416, 1, 512, 512), (200, 1, 512, 512), the slice size is 512 × 512, the number of slices is different, and the processing is performed by a variable length sequence processing method: taking 500 as the length of CT data processing with different lengths, performing 0-complementing alignment on less than 500 frames, and finally processing the CT with different lengths to obtain data of (500, 1, 512, 512) formats of features to be extracted;
a feature extraction module: selecting a 2d feature extraction model for the data processed by the variable-length sequence processing mode, inputting (500, 1, 512, 512) obtained by preprocessing into the 2d feature extraction model, and obtaining a (500, 4096) format feature vector;
a model training module: inputting the feature vectors (N, 500, 4096) with the number of the training sets N into a recurrent neural network for training to obtain a final deep learning model.
While specific embodiments of the present invention have been described in detail above, it will be understood that modifications may be made thereto without departing from the spirit of the invention. It is intended that the following claims cover such modifications as fall within the true scope and spirit of the invention.
Claims (2)
1. A deep learning method for analyzing high-dimensional medical data is characterized by comprising the following steps:
s1: reading high-dimensional medical data, selecting one of a fixed-length interval extraction mode, a fixed-length overlapping extraction mode and a variable-length sequence processing mode as a preprocessing mode according to the attributes of the high-dimensional medical data by analyzing the attributes of the high-dimensional medical data, and preprocessing the high-dimensional medical data by using the selected preprocessing mode;
s2: selecting different feature extraction models according to the selected preprocessing mode, and performing feature extraction on the preprocessed data through the selected feature extraction models to obtain corresponding feature vectors;
s3: training the corresponding feature vectors to obtain a final deep learning model for medical data analysis;
selecting a 2d convolutional neural network or a 2d cyclic neural network for the data format obtained by processing the fixed-length interval extraction mode and the variable-length sequence processing mode, and selecting a 3d convolutional neural network or a 3d cyclic neural network for the data format obtained by processing the fixed-length overlapping extraction mode;
the fixed-length interval extraction mode comprises the following steps: taking the middle layer of the data with different lengths as a reference, respectively selecting the data to two sides at a certain interval, and finally forming a data format with a certain length for a subsequent feature extraction model;
the fixed-length overlapping extraction mode comprises the following steps: for data with different lengths, a certain number of frame numbers are extracted from one end of the data to serve as a first group of data, then the same number of frame numbers are extracted to serve as a second group of data, the two groups of data are partially overlapped, and finally a data format with a certain length is formed for a subsequent feature extraction model;
the variable length sequence processing mode comprises the following steps: and for data with different lengths, adding the data into a predetermined data format with a certain length, and if the length of the data is smaller than the fixed length, adding 0 to the subsequent data for alignment so as to be used by a subsequent feature extraction model.
2. A deep learning apparatus that analyzes high-dimensional medical data, characterized by comprising:
the data preprocessing module reads the high-dimensional medical data, selects one of a fixed-length interval extraction mode, a fixed-length overlapping extraction mode and a variable-length sequence processing mode according to the attributes of the high-dimensional medical data as a preprocessing mode by analyzing the attributes of the high-dimensional medical data, and preprocesses the high-dimensional medical data by using the selected preprocessing mode;
the feature extraction module is used for receiving the preprocessed data, selecting different feature extraction models according to the selected preprocessing mode, and performing feature extraction on the preprocessed data through the selected feature extraction models to obtain corresponding feature vectors;
the model training module receives and trains the corresponding characteristic vectors to obtain a final deep learning model for medical data analysis;
selecting a 2d convolutional neural network or a 2d cyclic neural network for the data format obtained by processing the fixed-length interval extraction mode and the variable-length sequence processing mode, and selecting a 3d convolutional neural network or a 3d cyclic neural network for the data format obtained by processing the fixed-length overlapping extraction mode;
the fixed-length interval extraction mode comprises the following steps: taking the middle layer of the data with different lengths as a reference, respectively selecting the data to two sides at a certain interval, and finally forming a data format with a certain length for a subsequent feature extraction model;
the fixed-length overlapping extraction mode comprises the following steps: for data with different lengths, a certain number of frame numbers are extracted from one end of the data to serve as a first group of data, then the same number of frame numbers are extracted to serve as a second group of data, the two groups of data are partially overlapped, and finally a data format with a certain length is formed for a subsequent feature extraction model;
the variable length sequence processing mode comprises the following steps: and for data with different lengths, adding the data into a predetermined data format with a certain length, and if the length of the data is smaller than the fixed length, adding 0 to the subsequent data for alignment so as to be used by a subsequent feature extraction model.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611122716.4A CN108198625B (en) | 2016-12-08 | 2016-12-08 | Deep learning method and device for analyzing high-dimensional medical data |
| PCT/CN2017/112606 WO2018103538A1 (en) | 2016-12-08 | 2017-11-23 | Deep learning method and device for analysis of high-dimensional medical data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611122716.4A CN108198625B (en) | 2016-12-08 | 2016-12-08 | Deep learning method and device for analyzing high-dimensional medical data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108198625A CN108198625A (en) | 2018-06-22 |
| CN108198625B true CN108198625B (en) | 2021-07-20 |
Family
ID=62490657
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611122716.4A Active CN108198625B (en) | 2016-12-08 | 2016-12-08 | Deep learning method and device for analyzing high-dimensional medical data |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN108198625B (en) |
| WO (1) | WO2018103538A1 (en) |
Families Citing this family (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108960422B (en) * | 2018-06-19 | 2022-04-15 | 河南工业大学 | Width learning method based on principal component analysis |
| CN109300531A (en) * | 2018-08-24 | 2019-02-01 | 深圳大学 | A kind of cerebral disease method of early diagnosis and device |
| CN109448855A (en) * | 2018-09-17 | 2019-03-08 | 大连大学 | A kind of diabetes glucose prediction technique based on CNN and Model Fusion |
| CN109620203A (en) * | 2018-12-18 | 2019-04-16 | 安徽心之声医疗科技有限公司 | A kind of electrocardiosignal characteristic automatic extraction method based on one-dimensional convolutional neural networks |
| CN109902734A (en) * | 2019-02-22 | 2019-06-18 | 中国石油大学(华东) | A kind of cholelithiasis CT medical image method for quickly identifying based on YOLO convolutional neural networks |
| CN110222087B (en) * | 2019-05-15 | 2023-10-17 | 平安科技（深圳）有限公司 | Feature extraction method, device and computer readable storage medium |
| CN110751183A (en) * | 2019-09-24 | 2020-02-04 | 东软集团股份有限公司 | Image data classification model generation method, image data classification method and device |
| CN110826908A (en) * | 2019-11-05 | 2020-02-21 | 北京推想科技有限公司 | Evaluation method and device for artificial intelligent prediction, storage medium and electronic equipment |
| CN111243658B (en) * | 2020-01-07 | 2022-07-22 | 西南大学 | Biomolecular network construction and optimization method based on deep learning |
| CN112330644A (en) * | 2020-11-11 | 2021-02-05 | 复旦大学 | Medical image diagnosis system based on deep learning |
| CN113156243A (en) * | 2021-04-09 | 2021-07-23 | 南方电网电动汽车服务有限公司 | Fault prediction method and prediction system |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104866727A (en) * | 2015-06-02 | 2015-08-26 | 陈宽 | Deep learning-based method for analyzing medical data and intelligent analyzer thereof |
| CN105718952A (en) * | 2016-01-22 | 2016-06-29 | 武汉科恩斯医疗科技有限公司 | Method for focus classification of sectional medical images by employing deep learning network |
| WO2016145675A1 (en) * | 2015-03-13 | 2016-09-22 | 中国科学院声学研究所 | Big data processing method for segment-based two-grade deep learning model |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH07203400A (en) * | 1993-10-15 | 1995-08-04 | Matsushita Electric Ind Co Ltd | Multimedia rendering marker and its usage |
| CN105760507B (en) * | 2016-02-23 | 2019-05-03 | 复旦大学 | Cross-module state topic relativity modeling method based on deep learning |
| CN106097335B (en) * | 2016-06-08 | 2019-01-25 | 安翰光电技术（武汉）有限公司 | Alimentary canal lesion image identification system and recognition methods |
| CN106096616A (en) * | 2016-06-08 | 2016-11-09 | 四川大学华西医院 | Magnetic resonance image feature extraction and classification method based on deep learning |
-
2016
- 2016-12-08 CN CN201611122716.4A patent/CN108198625B/en active Active
-
2017
- 2017-11-23 WO PCT/CN2017/112606 patent/WO2018103538A1/en active Application Filing
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2016145675A1 (en) * | 2015-03-13 | 2016-09-22 | 中国科学院声学研究所 | Big data processing method for segment-based two-grade deep learning model |
| CN104866727A (en) * | 2015-06-02 | 2015-08-26 | 陈宽 | Deep learning-based method for analyzing medical data and intelligent analyzer thereof |
| CN105718952A (en) * | 2016-01-22 | 2016-06-29 | 武汉科恩斯医疗科技有限公司 | Method for focus classification of sectional medical images by employing deep learning network |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108198625A (en) | 2018-06-22 |
| WO2018103538A1 (en) | 2018-06-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108198625B (en) | Deep learning method and device for analyzing high-dimensional medical data | |
| CN107025369B (en) | Method and device for performing conversion learning on medical images | |
| CN107273800B (en) | Attention mechanism-based motion recognition method for convolutional recurrent neural network | |
| CN111127482B (en) | CT image lung and trachea segmentation method and system based on deep learning | |
| CN110992351B (en) | sMRI image classification method and device based on multi-input convolution neural network | |
| Bing et al. | Medical image super resolution using improved generative adversarial networks | |
| CN111369565B (en) | Digital pathological image segmentation and classification method based on graph convolution network | |
| CN109222972B (en) | fMRI whole brain data classification method based on deep learning | |
| CN110363760B (en) | Computer system for recognizing medical images | |
| CN111539290B (en) | Video motion recognition method and device, electronic equipment and storage medium | |
| CN108921942B (en) | Method and device for 2D (two-dimensional) conversion of image into 3D (three-dimensional) | |
| WO2020029915A1 (en) | Artificial intelligence-based device and method for tongue image splitting in traditional chinese medicine, and storage medium | |
| CN108647599B (en) | Human behavior recognition method combining 3D (three-dimensional) jump layer connection and recurrent neural network | |
| WO2022127500A1 (en) | Multiple neural networks-based mri image segmentation method and apparatus, and device | |
| CN109300531A (en) | A kind of cerebral disease method of early diagnosis and device | |
| EP4016454A1 (en) | Three-dimensional edge detection method and apparatus, storage medium and computer device | |
| CN116542991A (en) | Network architecture for fracture image segmentation, training method and segmentation method thereof | |
| CN110956617A (en) | Brain nuclear magnetic resonance abnormal image visualization method based on circulation attention model | |
| CN112836602A (en) | Behavior recognition method, device, equipment and medium based on space-time feature fusion | |
| Nguyen et al. | Efficient brain tumor segmentation with dilated multi-fiber network and weighted bi-directional feature pyramid network | |
| CN109996085A (en) | Model training method, image processing method, device and electronic equipment | |
| Zhao et al. | Rethinking superpixel segmentation from biologically inspired mechanisms | |
| Zhang et al. | Learn to walk across ages: Motion augmented multi-age group gait video translation | |
| Lu et al. | Mixed Resolution Network with hierarchical motion modeling for efficient action recognition | |
| Qu et al. | LWSFACE: Light-Weighted Weakly-Supervised 3D Facial Texture Reconstruction From a Single Image |
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 | ||
| CB02 | Change of applicant information |
Address after: Room B401, floor 4, building 1, No. 12, Shangdi Information Road, Haidian District, Beijing 100085 Applicant after: Tuxiang Medical Technology Co., Ltd Address before: 100025 C, Chaoyang District International Ocean Center, Beijing, 807 Applicant before: Beijing Tuoxiang Technology Co.,Ltd. |
|
| CB02 | Change of applicant information | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |