CN111627531A - Medical image classification processing system based on artificial intelligence - Google Patents
Medical image classification processing system based on artificial intelligence Download PDFInfo
- Publication number
- CN111627531A CN111627531A CN202010487137.XA CN202010487137A CN111627531A CN 111627531 A CN111627531 A CN 111627531A CN 202010487137 A CN202010487137 A CN 202010487137A CN 111627531 A CN111627531 A CN 111627531A
- Authority
- CN
- China
- Prior art keywords
- image data
- database
- inspection
- data
- 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.)
- Pending
Links
Classifications
-
- 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
- G16H30/00—ICT specially adapted for the handling or processing of medical images
- G16H30/20—ICT specially adapted for the handling or processing of medical images for handling medical images, e.g. DICOM, HL7 or PACS
-
- 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
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- 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
Abstract
The invention discloses a medical image classification processing system based on artificial intelligence, and mainly relates to the field of medical image classification. The system comprises a database, a database and a database server, wherein the database is configured in a PACS server and is used for acquiring and storing sample image data and inspection image data, the inspection image data is acquired by shooting through an image acquisition terminal, and the sample image data is label information which has a diagnosis result and is used for manually marking medical record information; the server is configured on the background server and used for realizing data interaction with the database and data interaction with the client, and the client is configured on the terminal computer. According to the invention, abundant image data are fully utilized through unified formats and manual marks, and the database is in cloud, so that the capacity expansion and sharing of the data are realized, and the construction of an efficient and accurate medical picture classification model is facilitated.
Description
Technical Field
The invention relates to the field of medical image classification, in particular to a medical image classification processing system based on artificial intelligence.
Background
With the rapid development of modern medical imaging technology, medical images have become important auxiliary diagnosis and treatment technologies. However, as various medical imaging devices such as CT, MR, DSA, DR, and a large number of computer technologies are fused to affect diagnosis, effective utilization of medical image resources has become a very urgent issue. The hospital generates massive medical images every day, and if the image categories can be automatically labeled, the workload of doctors can be reduced to a great extent, and the use efficiency of the medical images can also be improved. Medical image classification has become a very urgent need.
Although conventional image classification methods based on content are mostly based on global features of images, such as color, texture, shape features, and the like, these methods have achieved good classification performance, but development of such classification methods is greatly limited due to single thought. Deep learning can bring tremendous changes to the medical field, however, developing training models requires a great deal of expertise. In recent years, a plurality of papers have been reported in the field of lancets, namely lancets and lancet digital health, and the papers are one of the important directions for the development of the medical field.
Since the first proposal in the 80's of the 20 th century, deep learning has rapidly progressed in the last 10 years, largely due to advances in graphics processing functions originally developed for video games and increasingly large sets of source data, and since 2012, deep learning has revolutionized various fields of computer vision, image processing, speech recognition, natural language translation, robotics, and even automotive autopilot. In 2015, the journal "scientific americans" (ranks deep learning as one of the annual "changing world ideas").
Currently, there are three major obstacles to the application of deep learning techniques in medical treatment.
First, accessing large, well-formatted and results-accurate data sets is a significant challenge, and despite the fact that large clinical data sets are open to many institutions around the world, the amount of data available in a form that is easy to process and calculate, as well as accurate clinical diagnostic results that can be used in learning tasks, is small. The global scope is seen to be huge in data amount, but is difficult to effectively utilize.
Secondly, the existing device has huge data volume of shooting results, high requirements on computer hardware, and difficulty in burden of a single mechanism, and meanwhile, each mechanism and hospital try to establish a model and a system thereof, but a regional barrier exists, so that effective sharing of medical resources cannot be realized, and accuracy of deep learning is hindered.
Disclosure of Invention
The invention aims to provide a medical image classification processing system based on artificial intelligence, which fully utilizes abundant image data through unified formats and artificial marks, realizes the capacity expansion and sharing of the data by cloud on a database, and is beneficial to constructing an efficient and accurate medical image classification model.
In order to achieve the purpose, the invention is realized by the following technical scheme:
an artificial intelligence based medical image classification processing system, comprising:
the database is configured on the PACS server and used for acquiring and storing sample image data and inspection image data, the inspection image data is acquired by shooting through an image acquisition terminal, and the sample image data is label information which has a diagnosis result and is used for manually marking medical record information;
the server side is configured on the background server and used for realizing data interaction with the database and data interaction with the client side, and the server side comprises
The first communication module is configured to establish communication with the database and the client respectively;
the import unit is configured to acquire sample image data and inspection image data in the database;
the training unit is configured to train the sample image data based on a convolutional neural network to obtain a classification model;
the identification unit is configured to identify the inspection image data by using the classification model, match corresponding label information and obtain the disease category and judgment probability judged by the system;
and the second communication unit is configured to receive the query request of the client and feed back the matched disease types and probabilities to the client.
The client is configured on the terminal computer and comprises:
the inquiry unit is configured to send the identity information of the patient and an inquiry request to the server;
and the acquisition unit is used for acquiring the disease category and the judgment probability from the server.
The database is in data connection with image acquisition equipment of a hospital and receives inspection image data shot by the equipment.
The test image data is composed of image data and text data, and the text data includes identity information of the patient and examination items.
The sample image data is obtained by the following method:
s1, manually screening medical image pictures with determined medical diagnosis, wherein each disease condition is matched with 1-10 ten thousand as original image data;
s2, manually marking original image data, wherein the marked content comprises first label information and second label information, the first label information is inspection information, the inspection information comprises inspection equipment, inspection time, inspection parts, inspection items and departments of the original image data, and the second label information comprises medical diagnosis results corresponding to the original image data to obtain labeled image data;
s3, carrying out image denoising and enhancement on the labeled image data to obtain labeled sample image data;
s4, the sample image data is imported into the database.
The database includes: a sample storage module: for storing sample image data;
an archiving storage module: the image acquisition terminal is used for storing the inspection image data obtained from the image acquisition terminal.
Compared with the prior art, the invention has the beneficial effects that:
the system can assist a doctor in judging and give a disease suggestion to the medical image shot by the patient, so that the efficiency of the doctor in diagnosing the state of an illness is improved, and reliable assistance is provided for the final judgment of the doctor.
The system uses the convolutional neural network to carry out deep learning on the marked typical image data, thereby constructing a reliable classification model and realizing effective assistance of modern medical treatment.
The system unifies and integrates medical image resources through manual marking of information and manual screening, so that the existing medical images and diagnosis results can be fed back to application, and a reasonable and accurate classification model is established.
The system avoids the limitation of mechanisms and regions by cloud data, realizes data expansion and sharing, and provides an effective platform for the progress of doctors and the development of image medicine.
Detailed Description
The invention will be further illustrated with reference to the following specific examples. It should be understood that these examples are for illustrative purposes only and are not intended to limit the scope of the present invention. Further, it should be understood that various changes or modifications of the present invention may be made by those skilled in the art after reading the teaching of the present invention, and these equivalents also fall within the scope of the present application.
The instruments, reagents, materials and the like used in the following examples are conventional instruments, reagents, materials and the like in the prior art and are commercially available in a normal manner unless otherwise specified. Unless otherwise specified, the experimental methods, detection methods, and the like described in the following examples are conventional experimental methods, detection methods, and the like in the prior art.
Example (b): medical image classification processing system based on artificial intelligence
The classification system mainly takes deep learning as a classification basis and mainly comprises a server, a database and a client,
1) the database is configured in a PACS server, is in data connection with an image acquisition terminal (such as CT, nuclear magnetic resonance, ultrasonic examination and the like) of a hospital, receives and stores inspection image data, uses an s3 storage system as a file system of the image data to store mass image data, and uses RDS as an image database to store metadata information of the image data.
The problems that the data volume of the existing medical image data is large, and the requirement of local storage on hardware is high are solved. The S3 cloud platform provides an ultra-large data storage space to well solve the problem. The RDS cloud relational database provides instant backup and various fault-tolerant disaster-tolerant tools, and guarantees are provided for data safety.
The inspection image data is composed of image data and text data.
The method comprises the following steps:
a sample storage module: used for storing sample image data which are obtained from each open platform and have exact diagnosis results; the acquisition mode is manual screening and importing. The sample image data is obtained by processing the original image data, and the processing steps include:
s1, obtaining standard medical image pictures through the disclosed large medical image library platforms and systems, and manually selecting 1-10 million (depending on the complexity) original image data for each typical disease;
s2, labeling original image data by manual labeling, wherein the labeled content comprises first label information and second label information, the first label information is examination information, the examination information comprises examination equipment, examination time, examination parts, examination items and departments of the original image data, and the second label information comprises medical diagnosis results corresponding to the original image data; final to tagged image data;
s3, carrying out image denoising and enhancement on the labeled image data to obtain labeled sample image data;
s4, the sample image data is imported into the sample storage module.
An archiving storage module: the image acquisition terminal is used for storing inspection image data acquired from the image acquisition terminal, and the acquisition mode is communication connection and data import. Namely, the inspection result is imported into the database in real time, so that the data can be cloud.
2) Service terminal
And the server is configured at the background and used for realizing data interaction with the database and data interaction with the client.
The server comprises the following modules:
the first communication module is configured to establish communication with the database and the client respectively;
the import unit is configured to acquire image data in the database, and comprises sample image data imported into the sample storage module and inspection image data imported into the archiving storage module;
the training unit is configured to train the sample image data based on a convolutional neural network to obtain a classification model;
after processing, the image data can directly enter a convolutional neural network for training. In terms of neural network selection, we used the DenseNet at the leading edge of the current comparison. DenseNet is a convolutional neural network with dense connections. In the network, any two layers have direct connection, that is, the input of each layer of the network is the union of the outputs of all the previous layers, and the feature map learned by the layer is directly transmitted to all the next layers as input. The network has fewer parameters than a conventional convolutional network because it does not need to relearn the redundant feature map. And improves the transfer of information and gradients in the network, which makes the network easier to train.
The identification unit is configured to identify the inspection image data by using the classification model, match corresponding second label information and obtain the disease category and judgment probability judged by the system;
and the second communication unit is configured to receive the query request of the client and feed back the matched disease types and probabilities to the client.
3) Client terminal
The computer is configured on the terminal computer and can be a working computer of a doctor in a hospital.
The inquiry unit is used for inputting the identity information of the patient by a doctor in a mode of swiping a card (a medical insurance card or a treatment card) or manually inputting a mobile phone number or a medical insurance card number of the patient and sending the identity information and an inquiry request to the server;
the acquiring unit is used for acquiring the disease category and the judgment probability from the server;
the system can assist a doctor in judging and give a disease suggestion to the medical image shot by the patient, so that the efficiency of the doctor in diagnosing the state of an illness is improved, and reliable assistance is provided for the final judgment of the doctor.
The system uses the convolutional neural network to carry out deep learning on the marked typical image data, thereby constructing a reliable classification model and realizing effective assistance of modern medical treatment.
Based on the system, a deep combination module can be added, and the matched result, the subsequent tracking condition, the treatment means and the recovery condition are all brought into a database for storage and used as the basis and the precious resources for the study of doctors. Through network sharing and the capacity expansion storage mechanism of the cloud platform, the doctor is supported by big data of medical record data and diagnosis, and the doctor is helped to grow up quickly.
Claims (5)
1. A medical image classification processing system based on artificial intelligence, comprising:
the database is configured on the PACS server and used for acquiring and storing sample image data and inspection image data, the inspection image data is acquired by shooting through an image acquisition terminal, and the sample image data is label information which has a diagnosis result and is used for manually marking medical record information;
the server configured in the background server is used for realizing data interaction with the database and data interaction with the client, and comprises:
the first communication module is configured to establish communication with the database and the client respectively;
the import unit is configured to acquire sample image data and inspection image data in the database;
the training unit is configured to train the sample image data based on a convolutional neural network to obtain a classification model;
the identification unit is configured to identify the inspection image data by using the classification model, match corresponding label information and obtain the disease category and judgment probability judged by the system;
and the second communication unit is configured to receive the query request of the client and feed back the matched disease types and probabilities to the client.
The client is configured on the terminal computer and comprises:
the inquiry unit is configured to send the identity information of the patient and an inquiry request to the server;
and the acquisition unit is used for acquiring the disease category and the judgment probability from the server.
2. The artificial intelligence based medical image classification processing system according to claim 1, wherein the database is in data connection with an image acquisition device of a hospital, and receives inspection image data captured by the device.
3. The artificial intelligence based medical image classification processing system according to claim 1, wherein the verification image data is composed of image data and text data, the text data including identity information of a patient and examination items.
4. The artificial intelligence based medical image classification processing system according to claim 1, wherein the sample image data is obtained by:
s1, manually screening medical image pictures with determined medical diagnosis, wherein each disease condition is matched with 1-10 ten thousand as original image data;
s2, manually marking original image data, wherein the marked content comprises first label information and second label information, the first label information is inspection information, the inspection information comprises inspection equipment, inspection time, inspection parts, inspection items and departments of the original image data, and the second label information comprises medical diagnosis results corresponding to the original image data to obtain labeled image data;
s3, carrying out image denoising and enhancement on the labeled image data to obtain labeled sample image data;
s4, the sample image data is imported into the database.
5. The artificial intelligence based medical image classification processing system according to claim 1, wherein the database includes:
a sample storage module: for storing sample image data;
an archiving storage module: the image acquisition terminal is used for storing the inspection image data obtained from the image acquisition terminal.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010487137.XA CN111627531A (en) | 2020-06-02 | 2020-06-02 | Medical image classification processing system based on artificial intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010487137.XA CN111627531A (en) | 2020-06-02 | 2020-06-02 | Medical image classification processing system based on artificial intelligence |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111627531A true CN111627531A (en) | 2020-09-04 |
Family
ID=72272007
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010487137.XA Pending CN111627531A (en) | 2020-06-02 | 2020-06-02 | Medical image classification processing system based on artificial intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111627531A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111931717A (en) * | 2020-09-22 | 2020-11-13 | 平安科技（深圳）有限公司 | Semantic and image recognition-based electrocardiogram information extraction method and device |
| CN112201327A (en) * | 2020-10-09 | 2021-01-08 | 福建智康云医疗科技有限公司 | Remote medical image system based on cloud data |
| CN112259198A (en) * | 2020-10-21 | 2021-01-22 | 上海市同仁医院 | Medical image data management system and method |
| CN112447280A (en) * | 2021-01-08 | 2021-03-05 | 深圳坐标软件集团有限公司 | Intelligent medical system for medical image information management |
| CN113096779A (en) * | 2021-04-23 | 2021-07-09 | 深圳市龙岗区第三人民医院 | Medical image cloud image system based on distributed CT terminal |
| CN113496774A (en) * | 2021-07-22 | 2021-10-12 | 河北北方学院 | Digital medical information management method and related device |
| CN116741350A (en) * | 2023-08-16 | 2023-09-12 | 枣庄市山亭区妇幼保健院 | File management system for hospital X-ray images |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103116626A (en) * | 2013-01-31 | 2013-05-22 | 张粤新 | Matching method of medical image |
| CN103164615A (en) * | 2013-02-01 | 2013-06-19 | 尹新平 | Skin image analysis management system |
| CN110909780A (en) * | 2019-11-14 | 2020-03-24 | 腾讯科技（深圳）有限公司 | Image recognition model training and image recognition method, device and system |
| CN110993064A (en) * | 2019-11-05 | 2020-04-10 | 北京邮电大学 | Deep learning-oriented medical image labeling method and device |
-
2020
- 2020-06-02 CN CN202010487137.XA patent/CN111627531A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103116626A (en) * | 2013-01-31 | 2013-05-22 | 张粤新 | Matching method of medical image |
| CN103164615A (en) * | 2013-02-01 | 2013-06-19 | 尹新平 | Skin image analysis management system |
| CN110993064A (en) * | 2019-11-05 | 2020-04-10 | 北京邮电大学 | Deep learning-oriented medical image labeling method and device |
| CN110909780A (en) * | 2019-11-14 | 2020-03-24 | 腾讯科技（深圳）有限公司 | Image recognition model training and image recognition method, device and system |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111931717A (en) * | 2020-09-22 | 2020-11-13 | 平安科技（深圳）有限公司 | Semantic and image recognition-based electrocardiogram information extraction method and device |
| CN111931717B (en) * | 2020-09-22 | 2021-01-26 | 平安科技（深圳）有限公司 | Semantic and image recognition-based electrocardiogram information extraction method and device |
| CN112201327A (en) * | 2020-10-09 | 2021-01-08 | 福建智康云医疗科技有限公司 | Remote medical image system based on cloud data |
| CN112259198A (en) * | 2020-10-21 | 2021-01-22 | 上海市同仁医院 | Medical image data management system and method |
| CN112447280A (en) * | 2021-01-08 | 2021-03-05 | 深圳坐标软件集团有限公司 | Intelligent medical system for medical image information management |
| CN113096779A (en) * | 2021-04-23 | 2021-07-09 | 深圳市龙岗区第三人民医院 | Medical image cloud image system based on distributed CT terminal |
| CN113496774A (en) * | 2021-07-22 | 2021-10-12 | 河北北方学院 | Digital medical information management method and related device |
| CN116741350A (en) * | 2023-08-16 | 2023-09-12 | 枣庄市山亭区妇幼保健院 | File management system for hospital X-ray images |
| CN116741350B (en) * | 2023-08-16 | 2023-10-31 | 枣庄市山亭区妇幼保健院 | File management system for hospital X-ray images |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111627531A (en) | Medical image classification processing system based on artificial intelligence | |
| CN112949786A (en) | Data classification identification method, device, equipment and readable storage medium | |
| Gauriau et al. | Using DICOM metadata for radiological image series categorization: a feasibility study on large clinical brain MRI datasets | |
| CN112766377B (en) | Left ventricle magnetic resonance image intelligent classification method, device, equipment and medium | |
| CN111291825A (en) | Focus classification model training method and device, computer equipment and storage medium | |
| CN113821668A (en) | Data classification identification method, device, equipment and readable storage medium | |
| WO2022134357A1 (en) | Triage data processing method and apparatus, and computer device and storage medium | |
| CN113656706A (en) | Information pushing method and device based on multi-mode deep learning model | |
| CN115908800A (en) | Medical image segmentation method | |
| CN116469545A (en) | Method, device and equipment for carrying out auxiliary disease prediction by using medical image | |
| CN117237351B (en) | Ultrasonic image analysis method and related device | |
| Qin et al. | Residual block-based multi-label classification and localization network with integral regression for vertebrae labeling | |
| CN111898528B (en) | Data processing method, device, computer readable medium and electronic equipment | |
| CN112397178A (en) | Medical image classification processing system based on artificial intelligence | |
| CN115588492A (en) | Diagnosis guiding method and system based on artificial intelligence | |
| WO2023029348A1 (en) | Image instance labeling method based on artificial intelligence, and related device | |
| CN113707304B (en) | Triage data processing method, triage data processing device, triage data processing equipment and storage medium | |
| CN114332574A (en) | Image processing method, device, equipment and storage medium | |
| Giordano et al. | An AI-based framework for supporting large scale automated analysis of video capsule endoscopy | |
| CN113793677A (en) | Electronic medical record management method and device, storage medium and electronic equipment | |
| Jin et al. | Content and semantic context based image retrieval for medical image grid | |
| CN115700826A (en) | Receipt processing method, receipt display method, receipt processing device, receipt display device, computer equipment and storage medium | |
| Peng et al. | Multi-view weighted feature fusion using cnn for pneumonia detection on chest x-rays | |
| Jianjian et al. | MCSC-UTNet: Honeycomb lung segmentation algorithm based on Separable Vision Transformer and context feature fusion | |
| CN111105871A (en) | Medical auxiliary diagnosis method and system |
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 |