CN110675378B - Image identification method and system for stability of spinal metastasis tumor - Google Patents
Image identification method and system for stability of spinal metastasis tumor Download PDFInfo
- Publication number
- CN110675378B CN110675378B CN201910898387.XA CN201910898387A CN110675378B CN 110675378 B CN110675378 B CN 110675378B CN 201910898387 A CN201910898387 A CN 201910898387A CN 110675378 B CN110675378 B CN 110675378B
- Authority
- CN
- China
- Prior art keywords
- neural network
- module
- network model
- training
- model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- 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
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- 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]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30008—Bone
- G06T2207/30012—Spine; Backbone
-
- 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30096—Tumor; Lesion
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Quality & Reliability (AREA)
- Artificial Intelligence (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Medical Informatics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Apparatus For Radiation Diagnosis (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses an image identification method and system for stability of spinal metastasis tumor. The image identification method and the system for the stability of the spinal metastasis tumor are adopted to collect an image data set, and the data set is divided into a training set and a testing set; labeling data in the image data set; splitting a plurality of parts which can be scored according to image information in the image data set into a plurality of subtasks, and establishing a deep neural network model for each subtask; training the neural network model by using the data of the training set to obtain a system model; and testing the system model by using the data of the test set. The image recognition method and the system for the stability of the spinal metastasis tumor can establish a deep neural network by using an artificial intelligence technology, and obtain a system model for judging the stability of the spinal metastasis tumor through training of a spinal metastasis tumor CT image database.
Description
Technical Field
The invention relates to the field of image recognition of spinal metastasis tumors, in particular to an image recognition method and system for stability of spinal metastasis tumors.
Background
At present, the stability of the spinal metastatic tumor and the stability judgment of the spinal metastatic tumor usually adopt the manual identification of CT images one by one, which is a huge test for the experience and the technical level of doctors, and simultaneously consume huge labor cost, time cost, technical cost and equipment cost. Therefore, the technology that the artificial intelligence technology can be utilized to establish the deep neural network, the artificial intelligence can be used for intelligently identifying the spine CT image of the spine metastasis tumor through training of the spine CT image database, and the automatic judgment of the stability of the spine metastasis tumor is very important.
Aiming at the problems in the prior art, the method and the system for identifying the stability of the spinal metastasis tumor have important significance.
Disclosure of Invention
In order to solve the above problems, the present invention provides an image recognition method and system for spinal metastatic tumor stability.
In order to achieve the above object, the present invention provides an image recognition method for spinal metastatic tumor stability, comprising the following steps: collecting an image data set, and dividing the data set into a training set and a test set; labeling data in the image data set; splitting a plurality of parts which can be scored according to image information in the image data set into a plurality of subtasks, and establishing a deep neural network model for each subtask; training the neural network model by using the data of the training set to obtain a system model; testing the system model using the data of the test set;
further, the dividing a plurality of parts which can be scored according to image information in the image data set into a plurality of subtasks, and establishing a deep neural network model for each subtask specifically includes the following steps: iterating the data in the image data set, and correcting errors caused by labeling; suppressing overfitting of the neural network model; guiding the attention of the neural network model and improving the accuracy of the neural network model; improving the collapse discrimination of the neural network model based on a collapse classification method of ResNet;
further, the training of the neural network model using the data of the training set to obtain a system model specifically includes the following steps: optimizing the neural network model using a stochastic gradient descent or Adam optimizer; initializing the neural network model using pre-trained model parameters; accelerating convergence of the neural network model;
further, iterating the data in the image data set, and correcting an error caused by labeling specifically includes: iterating the data in the image data set by a weak supervision segmentation method based on a Mask RCNN network and a self-training method, and correcting errors caused by labeling;
further, the inhibiting of the overfitting of the neural network model specifically includes: based on a classification method of a DenseNet neural network and a multitask learning method and a classification method of the damage condition of the posterolateral side of a vertebral body, the overfitting of the neural network model is inhibited according to the detection results of the bone mass and the damage condition of the posterolateral side which are simultaneously output after the cross section;
further, the inhibiting of the overfitting of the neural network model specifically includes: based on a classification method of a DenseNet neural network and a multitask learning method and a classification method of the damage condition of the posterolateral side of a vertebral body, the overfitting of the neural network model is inhibited according to the detection results of the bone mass and the damage condition of the posterolateral side which are simultaneously output after the cross section;
the invention also provides an image recognition system for the stability of the spinal metastasis tumor, which comprises a data set storage module, a labeling module, a neural network model module, a system model module, a training module and a testing module, wherein the data set storage module is used for storing an image data set and a training set and a testing set which are distinguished by the image data set; the labeling module is used for labeling the data in the image data set; the neural network model module is used for storing the neural network model; the training module is used for training the neural network model to obtain a system model; the test module is used for testing the system model; the system model module is used for storing the system model;
further, the device also comprises an iteration module, a restraining module, a guiding force line module and a collapse classification module; the iteration module is used for iterating the data in the image data set and correcting errors caused by labeling; the inhibition module is used for inhibiting overfitting of the neural network model; the guiding force line module is used for guiding the attention of the neural network model and improving the accuracy of the neural network model; the collapse classification module is used for improving the collapse discrimination of the neural network model based on a collapse classification method of ResNet;
further, the device also comprises an optimization module, an initialization module and an accelerated convergence module; the optimization module is used for optimizing the neural network model by using a random gradient descent or an Adam optimizer; the initialization module is used for initializing the neural network model by using pre-trained model parameters; the accelerated convergence module is used for accelerating the convergence of the neural network model;
further, the iteration module specifically performs iteration on data in the image data set by a weak supervision segmentation method based on a Mask RCNN network and a self-training method, and corrects errors caused by labeling;
furthermore, the inhibition module is a classification method based on a DenseNet neural network and a multi-task learning method and a classification method of the damaged condition of the posterior and lateral side of the vertebral body, and inhibits overfitting of the neural network model according to the detection results of the quality of the bone and the damaged condition of the posterior and lateral side which are simultaneously output after the cross section.
The image recognition method and the system for the stability of the spinal column metastasis tumor can establish a deep neural network by using an artificial intelligence technology, obtain a system model for judging the stability of the spinal column metastasis tumor by training a CT image database of the spinal column metastasis tumor, and have the advantages that:
1. the diagnosis and evaluation time of clinical doctors in related departments such as oncology, imaging department, spinal surgery and the like can be obviously shortened, and the clinical workload is reduced. 2. The invention can be applied to clinical teaching work of medical students and training work of low-age funding doctors, and saves the manpower and material resources of hospitals. 3. The technical scheme of the invention has low cost and high efficiency, and can be conveniently arranged in a primary hospital to improve the diagnosis and treatment level of the primary hospital.
Drawings
Fig. 1 is a first flowchart of an image identification method for spinal metastatic tumor stability according to the present invention;
FIG. 2 is a second flowchart of an image recognition method for spinal metastatic tumor stability according to the present invention;
FIG. 3 is a third flowchart of an image recognition method for spinal metastatic tumor stability according to the present invention;
fig. 4 is a schematic structural diagram of an image recognition system for spinal metastatic tumor stability according to the present invention.
Detailed Description
The structure, operation, and the like of the present invention will be further described with reference to the accompanying drawings.
As shown in fig. 1, fig. 1 is a first flowchart of an image identification method for stability of spinal metastatic tumor according to the present invention, and the image identification method for stability of spinal metastatic tumor includes the following steps:
s1 collecting an image data set and dividing the data set into a training set and a testing set; in a preferred embodiment of the present invention, the image dataset is a spine CT image set of spine metastases;
s2 labeling the data in the image dataset; in a preferred embodiment of the present invention, the labeling is specifically labeling the data in the image data set correspondingly, and assigning a medical label;
s3, dividing a plurality of parts which can be scored according to image information in the image data set into a plurality of subtasks, and establishing a deep neural network model for each subtask;
s4, training the neural network model by using the data of the training set to obtain a system model;
s5 testing the system model using the data of the test set.
As shown in fig. 2, fig. 2 is a second flowchart of the image identification method for spine metastasis tumor stability according to the present invention, and the step S3 of splitting a plurality of parts that can be scored according to image information in the image data set into a plurality of subtasks and establishing a deep neural network model for each subtask specifically includes the following steps:
s31, iterating the data in the image data set, and correcting errors caused by labeling; in a preferred embodiment of the invention, iteration is performed on data in the image data set, specifically, in consideration of the fact that segmentation task labeling under strong supervision is difficult to obtain and time is consumed, a weak supervision training method is adopted, a backbone vertebral body weak supervision segmentation method based on a Mask RCNN network and a self-training method is adopted, each patient only needs to select a middle sagittal plane and label the positions of four corners of each vertebral body, then iteration is performed by combining the error correction capability and self-supervision of the Mask RCNN, errors caused by weak labeling are gradually corrected, and 3D spine segmentation is finally completed;
s32 suppressing overfitting of the neural network model; in the preferred embodiment of the invention, the invention specifically relates to a bone quality classification method and a vertebral body posterolateral injury classification method based on a DenseNet neural network and a multitask learning method. The bone mass and the damage condition of the back and the outer side of the vertebral body are both shown from the cross section, so the idea of multi-task learning is adopted, the detection results of the bone mass and the damage condition of the back and the outer side are simultaneously output after the cross section is input, a plurality of tasks are mutually promoted, and the overfitting of a network is effectively inhibited;
s33 guiding the attention of the neural network model and improving the accuracy of the neural network model; in the preferred embodiment of the invention, the force line classification method is specifically based on expert knowledge guidance. The force line classification of the spine relates to the characteristics of fine-grained classification, the network is difficult to capture due to the slight difference between positive and negative samples, and the model is guided to pay attention to the two sides of the spine by introducing expert knowledge of doctors, namely the main distinguishing areas of the force line problem of the spine are positioned at the front edge and the rear edge of the spine, so that the accuracy of the model is improved;
s34 improves the collapse discrimination of the neural network model based on the collapse classification method of ResNet. In a preferred embodiment of the invention, in particular a collapse classification method based on ResNet. Using ResNet directly as the basic model for classification, since spine collapse is often difficult to diagnose by a single spine, in many cases requiring comparison with the volumes of the upper and lower spines, ResNet was modified to a 3-channel input, with labeling of the middle spine as a supervision, and L2 norms for the middle spine and the two adjacent spines, respectively, as constraints, making < 50% and > 50% collapse easier to distinguish.
As shown in fig. 3, fig. 3 is a third flow chart of the image identification method for spinal metastatic tumor stability according to the present invention;
the S4 training the neural network model using the data of the training set to obtain a system model, specifically including the following steps:
s41 optimizing the neural network model using a stochastic gradient descent or Adam optimizer;
s42 initializing the neural network model using the pre-trained model parameters;
s43 accelerates convergence of the neural network model, avoiding falling into a locally optimal solution.
As shown in fig. 4, fig. 4 is a schematic structural diagram of an image recognition system for stability of a spinal metastatic tumor, which includes a data set storage module 1, an annotation module 2, an iteration module 7, a suppression module 8, a guiding force line module 9, a collapse classification module 10, a neural network model module 3, a system model module 4, a training module 6 and a test module 5, where the data set storage module 1 is configured to store an image data set and a training set and a test set distinguished from the image data set; the labeling module 2 is configured to label data in the image dataset; the neural network model module 3 is used for storing the neural network model; the training module 6 is used for training the neural network model to obtain a system model; the test module 5 is used for testing the system model; the system model module 4 is used for storing the system model; the iteration module 7 is configured to iterate data in the image data set and correct errors caused by labeling, and in a preferred embodiment of the present invention, the iteration module 7 specifically iterates data in the image data set by using a Mask RCNN network and a weak supervised segmentation method based on a self-training method to correct errors caused by labeling; the inhibition module 8 is configured to inhibit overfitting of the neural network model, and in a preferred embodiment of the present invention, the inhibition module 8 is specifically configured to output a detection result of bone mass and a detection result of a posterior-lateral side damaged condition simultaneously according to a cross section based on a classification method of a DenseNet neural network and a multitask learning method and a classification method of a posterior-lateral side damaged condition of a vertebral body, so as to inhibit overfitting of the neural network model; the guiding force line module 9 is used for guiding the attention of the neural network model and improving the accuracy of the neural network model; the collapse classification module 10 is used for improving the collapse discrimination of the neural network model based on a collapse classification method of ResNet;
the system also comprises an optimization module 11, an initialization module 12 and an accelerated convergence module 13; the optimization module 11 is configured to optimize the neural network model using a stochastic gradient descent or Adam optimizer; the initialization module 12 is configured to initialize the neural network model using the pre-trained model parameters; the accelerated convergence module 13 is configured to accelerate convergence of the neural network model.
In a first preferred embodiment of the present invention, the image recognition method for the stability of the spinal metastatic tumor comprises the following steps:
s1, collecting an image data set, wherein the image data set is a spine CT image of 800 cases of spine metastasis tumors, and dividing the data set into a training set and a testing set;
s2 labeling the data in the image data set correspondingly and assigning a medical label;
s3, selecting 5 parts which can be scored according to image information from the image data set and splitting the parts into 5 subtasks respectively, and establishing a deep neural network model for each subtask;
s4, training a neural network by using the training set data to obtain a system model;
s5 testing the system model using the data of the test set.
The foregoing is merely illustrative of the present invention, and it will be appreciated by those skilled in the art that various modifications may be made without departing from the principles of the invention, and the scope of the invention is to be determined accordingly.
Claims (6)
1. An image identification method for stability of a spinal metastasis tumor is characterized by comprising the following steps:
collecting an image data set, and dividing the data set into a training set and a test set;
labeling data in the image data set;
splitting a plurality of parts which can be scored according to image information in the image data set into a plurality of subtasks, and establishing a deep neural network model for each subtask; the method specifically comprises the following steps: iterating the data in the image data set, and correcting errors caused by labeling; suppressing overfitting of the neural network model; guiding the attention of the neural network model and improving the accuracy of the neural network model; improving the collapse discrimination of the neural network model based on a collapse classification method of ResNet; wherein the inhibiting of the overfitting of the neural network model specifically comprises: based on a classification method of a DenseNet neural network and a multitask learning method and a classification method of the damage condition of the posterolateral side of a vertebral body, the overfitting of the neural network model is inhibited according to the detection results of the bone mass and the damage condition of the posterolateral side which are simultaneously output after the cross section;
training the neural network model by using the data of the training set to obtain a system model;
and testing the system model by using the data of the test set.
2. The image recognition method for spinal metastatic tumor stability according to claim 1, wherein the training of the neural network model using the data of the training set to obtain a system model comprises the following steps:
optimizing the neural network model using a stochastic gradient descent or Adam optimizer;
initializing the neural network model using pre-trained model parameters;
accelerating convergence of the neural network model.
3. The image recognition method for spine metastasis tumor stability according to claim 1, wherein the data in the image dataset are iterated to correct errors caused by labeling, specifically: and iterating the data in the image data set by a weak supervision segmentation method based on a Mask RCNN network and a self-training method, and correcting errors caused by labeling.
4. An image recognition system for stability of a spinal metastasis tumor is characterized by comprising a data set storage module, a labeling module, a neural network model module, a system model module, a training module and a testing module, wherein the data set storage module is used for storing an image data set and a training set and a testing set which are distinguished by the image data set; the labeling module is used for labeling the data in the image data set; the neural network model module is used for storing the neural network model; the training module is used for training the neural network model to obtain a system model; the test module is used for testing the system model; the system model module is used for storing the system model;
the image recognition system also comprises an iteration module, a suppression module, a guide force line module and a collapse classification module; the iteration module is used for iterating the data in the image data set and correcting errors caused by labeling; the inhibition module is used for inhibiting overfitting of the neural network model; the guiding line-of-force module is used for guiding the attention of the neural network model and improving the accuracy of the neural network model; the collapse classification module is used for improving the collapse discrimination of the neural network model based on a collapse classification method of ResNet; the inhibition module is a classification method based on a DenseNet neural network and a multitask learning method and a classification method of the damage condition of the posterior and lateral side of the vertebral body, and inhibits overfitting of the neural network model according to the detection results of the bone mass and the damage condition of the posterior and lateral side of the posterior section.
5. The image recognition system for spinal metastatic tumor stability according to claim 4, further comprising an optimization module, an initialization module and an accelerated convergence module; the optimization module is used for optimizing the neural network model by using a random gradient descent or an Adam optimizer; the initialization module is used for initializing the neural network model by using pre-trained model parameters; the accelerated convergence module is used for accelerating convergence of the neural network model.
6. The image recognition system of spinal metastatic tumor stability according to claim 5, wherein the iteration module specifically iterates the data in the image dataset based on a Mask RCNN network and a weak supervised segmentation method of a self-training method, and corrects errors caused by labeling.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910898387.XA CN110675378B (en) | 2019-09-23 | 2019-09-23 | Image identification method and system for stability of spinal metastasis tumor |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910898387.XA CN110675378B (en) | 2019-09-23 | 2019-09-23 | Image identification method and system for stability of spinal metastasis tumor |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110675378A CN110675378A (en) | 2020-01-10 |
| CN110675378B true CN110675378B (en) | 2022-04-08 |
Family
ID=69077204
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910898387.XA Active CN110675378B (en) | 2019-09-23 | 2019-09-23 | Image identification method and system for stability of spinal metastasis tumor |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110675378B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111563897B (en) * | 2020-04-13 | 2024-01-05 | 北京理工大学 | Breast nuclear magnetic image tumor segmentation method and device based on weak supervision learning |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106530305A (en) * | 2016-09-23 | 2017-03-22 | 北京市商汤科技开发有限公司 | Semantic segmentation model training and image segmentation method and device, and calculating equipment |
| CN109190757A (en) * | 2018-08-07 | 2019-01-11 | 百度在线网络技术（北京）有限公司 | Task processing method, device, equipment and computer readable storage medium |
| CN109523532A (en) * | 2018-11-13 | 2019-03-26 | 腾讯科技（深圳）有限公司 | Image processing method, device, computer-readable medium and electronic equipment |
| CN110136828A (en) * | 2019-05-16 | 2019-08-16 | 杭州健培科技有限公司 | A method of medical image multitask auxiliary diagnosis is realized based on deep learning |
| CN110136133A (en) * | 2019-03-11 | 2019-08-16 | 嘉兴深拓科技有限公司 | A kind of brain tumor dividing method based on convolutional neural networks |
-
2019
- 2019-09-23 CN CN201910898387.XA patent/CN110675378B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106530305A (en) * | 2016-09-23 | 2017-03-22 | 北京市商汤科技开发有限公司 | Semantic segmentation model training and image segmentation method and device, and calculating equipment |
| CN109190757A (en) * | 2018-08-07 | 2019-01-11 | 百度在线网络技术（北京）有限公司 | Task processing method, device, equipment and computer readable storage medium |
| CN109523532A (en) * | 2018-11-13 | 2019-03-26 | 腾讯科技（深圳）有限公司 | Image processing method, device, computer-readable medium and electronic equipment |
| CN110136133A (en) * | 2019-03-11 | 2019-08-16 | 嘉兴深拓科技有限公司 | A kind of brain tumor dividing method based on convolutional neural networks |
| CN110136828A (en) * | 2019-05-16 | 2019-08-16 | 杭州健培科技有限公司 | A method of medical image multitask auxiliary diagnosis is realized based on deep learning |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110675378A (en) | 2020-01-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10452899B2 (en) | Unsupervised deep representation learning for fine-grained body part recognition | |
| CN110232383B (en) | Focus image recognition method and focus image recognition system based on deep learning model | |
| US11710233B2 (en) | Three-dimensional medical image analysis method and system for identification of vertebral fractures | |
| CN109859168B (en) | X-ray chest radiography image quality determination method and device | |
| US11275976B2 (en) | Medical image assessment with classification uncertainty | |
| EP3696727A1 (en) | Image analysis method and apparatus and computer program | |
| CN108292435A (en) | The spinal vertebrae positioning and segmentation based on study in 3D CT | |
| CN110175502A (en) | A kind of backbone Cobb angle measuring method, device, readable storage medium storing program for executing and terminal device | |
| EP2856197B1 (en) | Method and system for quantitative evaluation of image segmentation | |
| US10325068B2 (en) | Methods and apparatus to label radiology images | |
| CN103824090B (en) | Adaptive face low-level feature selection method and face attribute recognition method | |
| US20060252023A1 (en) | Methods for automatically identifying user selected answers on a test sheet | |
| US20220335600A1 (en) | Method, device, and storage medium for lesion segmentation and recist diameter prediction via click-driven attention and dual-path connection | |
| US20230005138A1 (en) | Lumbar spine annatomical annotation based on magnetic resonance images using artificial intelligence | |
| CN112802019B (en) | Leke typing method based on spine AIS image | |
| CN112037180B (en) | Chromosome segmentation method and device | |
| CN113393417A (en) | System and method for detecting laterality of medical images | |
| CN102831614A (en) | Sequential medical image quick segmentation method based on interactive dictionary migration | |
| CN110675378B (en) | Image identification method and system for stability of spinal metastasis tumor | |
| CN111667457A (en) | Medical image-based spine information automatic identification method, system, terminal and storage medium | |
| CN112561894B (en) | Intelligent electronic medical record generation method and system for CT image | |
| CN116228660A (en) | Method and device for detecting abnormal parts of chest film | |
| CN111459050B (en) | Intelligent simulation type nursing teaching system and teaching method based on dual-network interconnection | |
| CN112308820B (en) | Rib positioning method and device, computer equipment and storage medium | |
| CN113592834A (en) | X-ray chest radiography image quality determination method and device |
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 |