AI-based auxiliary diagnosis method

Abstract

The invention relates to clinical medicine, in particular to an AI-based auxiliary diagnosis method, which comprises the following steps of: collecting a large amount of clinical data, and processing the clinical data; key features are extracted from the clinical data, and the clinical data are labeled to construct a training set, a verification set and a test set; performing model training on the auxiliary diagnosis model by using the training set to obtain a trained auxiliary diagnosis model; deploying the trained auxiliary diagnosis model into a medical information system so as to process new clinical data in real time and give corresponding diagnosis suggestions; according to the technical scheme provided by the invention, the defects that the diagnosis accuracy and efficiency cannot be effectively improved and the workload of doctors cannot be reduced in the prior art can be effectively overcome.

Description

Technical Field

[0001] The present invention relates to clinical medicine, and in particular to an AI-based auxiliary diagnosis method. Background Art

[0002] The core of AI-based auxiliary diagnosis methods is to use artificial intelligence technology to help doctors diagnose diseases more accurately and quickly. Suppose a patient goes to the hospital for a chest X-ray. The traditional diagnosis method is for the doctor to personally view the X-ray image and judge whether there is an abnormality based on experience.

[0003] However, with AI-assisted diagnosis, X-ray images can be quickly analyzed, suspected problem areas can be marked, and possible diagnostic suggestions can even be given. In this way, doctors have more information and suggestions for reference when diagnosing, so that they can diagnose diseases more accurately and quickly. Therefore, the purpose of AI-based auxiliary diagnosis methods is to make AI a powerful assistant for doctors, improve the accuracy and efficiency of diagnosis through high-tech means, and allow patients to receive better medical services.

[0004] Most of the existing diagnostic methods rely on the subjective experience of doctors, which not only brings a heavy workload to doctors, but also cannot guarantee the accuracy and efficiency of diagnosis, making it impossible for patients to receive better medical services, resulting in low patient satisfaction. Summary of the invention

[0005] 1. Technical issues to be solved

[0006] In view of the above-mentioned shortcomings of the prior art, the present invention provides an AI-based auxiliary diagnosis method, which can effectively overcome the defects of the prior art that it cannot effectively improve the accuracy and efficiency of diagnosis and cannot reduce the workload of doctors.

[0007] (II) Technical solution

[0008] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0009] An AI-based auxiliary diagnosis method comprises the following steps:

[0010] S1. Collect a large amount of clinical data and process the clinical data;

[0011] S2. Extract key features from clinical data and annotate clinical data to construct training sets, validation sets, and test sets;

[0012] S3, using the training set to train the auxiliary diagnosis model to obtain a trained auxiliary diagnosis model;

[0013] S4. Deploy the trained auxiliary diagnosis model into the medical information system to process new clinical data in real time and provide corresponding diagnostic recommendations.

[0014] Preferably, S1 collects a large amount of clinical data and processes the clinical data, including:

[0015] Collect a large amount of clinical data, and perform data cleaning, data format standardization, and data anonymization on the clinical data;

[0016] Among them, clinical data include past patient medical history, electronic medical records, medical images, laboratory test results and physical examination results.

[0017] Preferably, S2 extracts key features from clinical data and annotates the clinical data to construct a training set, a validation set, and a test set, including:

[0018] S21. Extract key features that are helpful for disease diagnosis from clinical data, and give correct diagnosis results for clinical data to annotate clinical data;

[0019] S22, dividing the annotated clinical data into a training set, a validation set, and a test set according to a preset ratio, and generating a structured file containing the clinical data and the annotated results;

[0020] Among them, key features include specific edges, shapes, and textures detected in medical images.

[0021] Preferably, in S3, the auxiliary diagnosis model is trained using the training set to obtain a trained auxiliary diagnosis model, including:

[0022] Select a suitable machine learning model to build an auxiliary diagnosis model, and use the training set to train the auxiliary diagnosis model, so as to guide the auxiliary diagnosis model to learn the mapping relationship between clinical data and diagnostic results through model training, and obtain a trained auxiliary diagnosis model.

[0023] Preferably, the method of using the training set to perform model training on the auxiliary diagnosis model includes:

[0024] S31, setting the loss function and optimizer of the auxiliary diagnosis model, and inputting the training set into the auxiliary diagnosis model for model training;

[0025] S32, calculating the loss value based on the loss function, and the optimizer updates the model parameters according to the loss value and the network gradient information;

[0026] S33. If the loss value is less than the preset threshold, the model training is completed, and the current auxiliary diagnosis model is the trained auxiliary diagnosis model. Otherwise, return to S31 and continue to use the training set for model training.

[0027] Preferably, when the auxiliary diagnosis model is trained on the training set, the generalization ability of the model is evaluated by observing the performance on the validation set, and the hyperparameters and structure of the model are tuned;

[0028] After obtaining the trained auxiliary diagnosis model, the performance of the model is evaluated using a test set.

[0029] Preferably, in S4, the trained auxiliary diagnosis model is deployed to the medical information system to process new clinical data in real time and give corresponding diagnostic suggestions, including:

[0030] Conduct validation tests on the auxiliary diagnosis model in the actual clinical environment to evaluate its performance in the actual clinical environment and further optimize the model based on the feedback information collected;

[0031] During actual application, the performance of the model is continuously monitored, and new clinical data is collected for continuous learning and optimization of the model. The model is regularly updated to reflect the latest medical research and clinical practice.

[0032] Preferably, in S4, the trained auxiliary diagnosis model is deployed to the medical information system to process new clinical data in real time and give corresponding diagnostic suggestions, including:

[0033] By designing an intuitive user-friendly interface, medical staff can interact with the system in a friendly manner and easily access the diagnostic suggestions provided by the auxiliary diagnosis model.

[0034] (III) Beneficial effects

[0035] Compared with the prior art, the AI-based auxiliary diagnosis method provided by the present invention has the following beneficial effects:

[0036] 1) Improve diagnostic accuracy: AI algorithms can learn from large amounts of clinical data and detect factors that humans may overlook, thus reducing misdiagnosis and missed diagnoses;

[0037] 2) Accelerate diagnosis: For complex medical imaging data, AI can quickly perform image recognition and analysis, greatly shortening the diagnosis time;

[0038] 3) Data-driven decision support: By analyzing massive amounts of clinical data, AI can provide personalized diagnostic recommendations based on evidence-based medicine;

[0039] 4) Continuous learning and adaptation: AI systems can regularly update their models by continuously collecting new clinical data to ensure diagnostic accuracy;

[0040] 5) Multimodal data analysis capabilities: AI systems can process and analyze different types of clinical data, including text, numbers, and images, to provide a more comprehensive diagnostic perspective;

[0041] 6) Reduce doctors’ workload: Automated diagnostic processes reduce doctors’ repetitive work, allowing them to spend more time focusing on patients’ treatment;

[0042] 7) Improved patient engagement: With AI-assisted diagnosis, doctors can more easily explain diagnosis results and treatment plans to patients, allowing patients to more actively participate in their own health management;

[0043] 8) Reduce medical costs: In the long run, AI-assisted diagnosis can help medical institutions save resources, reduce the costs of unnecessary examinations and treatments, and enable patients to receive high-quality medical services. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0045] Figure 1 It is a schematic diagram of the process of the present invention;

[0046] Figure 2 It is a detailed schematic diagram of the present invention. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0048] An AI-based auxiliary diagnosis method, such as Figure 1 and Figure 2 As shown, S1. Collect a large amount of clinical data and process the clinical data, including:

[0049] Collect a large amount of clinical data and perform data cleaning (to remove inaccurate or incomplete information), data format standardization (to ensure that different types of data can be used together), and data anonymization (to protect patient privacy);

[0050] Among them, clinical data include patients' past medical history, electronic medical records, medical images (such as X-ray images, CT, MRI, etc.), laboratory test results and physical examination results.

[0051] S2. Extract key features from clinical data and annotate clinical data to construct training sets, validation sets, and test sets, including:

[0052] S21. Extract key features that are helpful for disease diagnosis from clinical data, and give correct diagnosis results for clinical data to annotate clinical data;

[0053] S22, dividing the annotated clinical data into a training set, a validation set, and a test set according to a preset ratio, and generating a structured file containing the clinical data and the annotated results;

[0054] Among them, key features include specific edges, shapes, and textures detected in medical images.

[0055] S3. Use the training set to train the auxiliary diagnosis model to obtain a trained auxiliary diagnosis model, specifically including:

[0056] Select a suitable machine learning model to build an auxiliary diagnosis model, and use the training set to train the auxiliary diagnosis model, so as to guide the auxiliary diagnosis model to learn the mapping relationship between clinical data and diagnostic results through model training, and obtain a trained auxiliary diagnosis model.

[0057] Specifically, the auxiliary diagnosis model is trained using the training set, including:

[0058] S31, setting the loss function and optimizer of the auxiliary diagnosis model, and inputting the training set into the auxiliary diagnosis model for model training;

[0059] S32, calculating the loss value based on the loss function, and the optimizer updates the model parameters according to the loss value and the network gradient information;

[0060] S33. If the loss value is less than the preset threshold, the model training is completed, and the current auxiliary diagnosis model is the trained auxiliary diagnosis model. Otherwise, return to S31 and continue to use the training set for model training.

[0061] In the above technical solution, when the auxiliary diagnosis model is trained on the training set, the generalization ability of the model is evaluated by observing the performance on the validation set, and the hyperparameters and structure of the model are tuned;

[0062] After obtaining the trained auxiliary diagnosis model, the performance of the model is evaluated using the test set.

[0063] In the technical solution of this application, an auxiliary diagnosis model can be constructed based on a convolutional neural network (CNN). When using a test set to evaluate the performance of the model, it is necessary to check indicators such as accuracy and recall. When the model performance is poor, different key features can be selected or different machine learning models can be tried to ensure diagnostic accuracy.

[0064] In S4, the trained auxiliary diagnosis model is deployed to the medical information system to process new clinical data in real time and provide corresponding diagnostic suggestions, including:

[0065] Conduct validation tests on the auxiliary diagnosis model in the actual clinical environment to evaluate its performance in the actual clinical environment and further optimize the model based on the feedback information collected;

[0066] During actual application, the performance of the model is continuously monitored, and new clinical data is collected for continuous learning and optimization of the model. The model is regularly updated to reflect the latest medical research and clinical practice.

[0067] In S4, the trained auxiliary diagnosis model is deployed to the medical information system to process new clinical data in real time and provide corresponding diagnostic suggestions, including:

[0068] By designing an intuitive user-friendly interface, medical staff can interact with the system in a friendly manner and easily access the diagnostic suggestions provided by the auxiliary diagnosis model.

[0069] In the technical solution of this application, it should be ensured that the development and use of auxiliary diagnosis models comply with relevant medical regulations and ethical standards, especially when processing sensitive medical data, and a corresponding review mechanism should be established to ensure that the diagnostic suggestions provided by the auxiliary diagnosis model will not mislead the doctor's judgment.

[0070] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An AI-based auxiliary diagnosis method, characterized in that: The following steps are involved: S1. Collect a large amount of clinical data and process the clinical data; S2. Extract key features from clinical data and annotate clinical data to construct training sets, validation sets, and test sets; S3, using the training set to train the auxiliary diagnosis model to obtain a trained auxiliary diagnosis model; S4. Deploy the trained auxiliary diagnosis model into the medical information system to process new clinical data in real time and provide corresponding diagnostic recommendations.

2. The AI-based auxiliary diagnosis method according to claim 1, characterized in that: S1 collects a large amount of clinical data and processes the clinical data, including: Collect a large amount of clinical data, and perform data cleaning, data format standardization, and data anonymization on the clinical data; Among them, clinical data include past patient medical history, electronic medical records, medical images, laboratory test results and physical examination results.

3. The AI-based auxiliary diagnosis method according to claim 1, characterized in that: S2 extracts key features from clinical data and annotates the clinical data to construct training sets, validation sets, and test sets, including: S21. Extract key features that are helpful for disease diagnosis from clinical data, and give correct diagnosis results for clinical data to annotate clinical data; S22, dividing the annotated clinical data into a training set, a validation set, and a test set according to a preset ratio, and generating a structured file containing the clinical data and the annotated results; Among them, key features include specific edges, shapes, and textures detected in medical images.

4. The AI-based auxiliary diagnosis method according to claim 1, characterized in that: In S3, the auxiliary diagnosis model is trained using the training set to obtain a trained auxiliary diagnosis model, including: Select a suitable machine learning model to build an auxiliary diagnosis model, and use the training set to train the auxiliary diagnosis model, so as to guide the auxiliary diagnosis model to learn the mapping relationship between clinical data and diagnostic results through model training, and obtain a trained auxiliary diagnosis model.

5. The AI-based auxiliary diagnosis method according to claim 4, characterized in that: The method of using the training set to perform model training on the auxiliary diagnosis model includes: S31, setting the loss function and optimizer of the auxiliary diagnosis model, and inputting the training set into the auxiliary diagnosis model for model training; S32, calculating the loss value based on the loss function, and the optimizer updates the model parameters according to the loss value and the network gradient information; S33. If the loss value is less than the preset threshold, the model training is completed, and the current auxiliary diagnosis model is the trained auxiliary diagnosis model. Otherwise, return to S31 and continue to use the training set for model training.

6. The AI-based auxiliary diagnosis method according to claim 5, characterized in that: When the auxiliary diagnosis model is trained on the training set, the generalization ability of the model is evaluated by observing the performance on the validation set, and the hyperparameters and structure of the model are tuned; After obtaining the trained auxiliary diagnosis model, the performance of the model is evaluated using a test set.

7. The AI-based auxiliary diagnosis method according to claim 1, characterized in that: In S4, the trained auxiliary diagnosis model is deployed to the medical information system to process new clinical data in real time and provide corresponding diagnostic suggestions, including: Conduct validation tests on the auxiliary diagnosis model in the actual clinical environment to evaluate its performance in the actual clinical environment and further optimize the model based on the collected feedback information; During actual application, the performance of the model is continuously monitored, and new clinical data is collected for continuous learning and optimization of the model. The model is regularly updated to reflect the latest medical research and clinical practice.

8. The AI-based auxiliary diagnosis method according to claim 1, characterized in that: In S4, the trained auxiliary diagnosis model is deployed to the medical information system to process new clinical data in real time and provide corresponding diagnostic suggestions, including: By designing an intuitive user-friendly interface, medical staff can interact with the system in a friendly manner and easily access the diagnostic suggestions provided by the auxiliary diagnosis model.

Images