Personalized intelligent medical service system based on electronic film

Abstract

The invention relates to intelligent medical treatment, in particular to a personalized intelligent medical service system based on an electronic film, which is characterized in that a server processes an electronic film image acquired by an image acquisition module through an image processing module, and constructs a focus recognition model by using a focus recognition model construction module; the server performs transfer learning by using the processed electronic film image through a focus recognition model training module, and inputs the to-be-recognized electronic film image of the patient into a trained focus recognition model by using the focus recognition module to obtain a focus recognition result; the server provides disease explanation for the patient through the disease explanation module according to the focus recognition result, and performs question-answer type disease interaction with the patient through the disease interaction module; according to the technical scheme provided by the invention, the defects that the focus in the electronic film image is difficult to accurately identify and friendly illness state interaction with the patient cannot be carried out in the prior art can be effectively overcome.

Description

Technical Field

[0001] The present invention relates to smart medical treatment, and in particular to a personalized smart medical treatment service system based on electronic film. Background Art

[0002] With the continuous advancement of medical technology and the growing needs of patients, the limitations of traditional film in the storage and transmission of medical images have become increasingly prominent. Traditional film is easily damaged, inconvenient to carry, and has limited information display, which cannot meet the needs of modern medicine for efficient, convenient, and accurate medical services. Therefore, electronic film has emerged as an information-based upgrade of traditional film. Based on mobile Internet and cloud storage technology, electronic film has the advantages of convenience, efficiency, and economy, and has been widely used in medical services.

[0003] At the same time, with the continuous development of technologies such as the Internet of Things, big data, and cloud computing, smart healthcare has gradually emerged, providing more intelligent and personalized solutions for medical services. However, in the process of developing personalized smart medical service systems based on electronic films, there are still technical challenges such as difficulty in accurately identifying lesions in electronic film images and inability to interact with patients in a friendly manner. Summary of the invention

[0004] 1. Technical issues to be solved

[0005] In view of the above-mentioned shortcomings of the prior art, the present invention provides a personalized intelligent medical service system based on electronic film, which can effectively overcome the defects of the prior art that it is difficult to accurately identify lesions in electronic film images and cannot interact with patients in a friendly manner about their condition.

[0006] (II) Technical solution

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0008] A personalized smart medical service system based on electronic film, comprising a server, wherein the server processes the electronic film image collected by the image collection module through the image processing module, and constructs a lesion recognition model using the lesion recognition model construction module, wherein the server performs transfer learning using the processed electronic film image through the lesion recognition model training module, and inputs the patient's electronic film image to be recognized into the trained lesion recognition model using the lesion recognition module to obtain a lesion recognition result;

[0009] The server provides the patient with a condition explanation based on the lesion identification result through the condition explanation module, and conducts a question-and-answer type condition interaction with the patient through the condition interaction module. The server provides the patient with a personalized treatment suggestion based on the lesion identification result through the treatment suggestion generation module;

[0010] The server receives the expert's lesion identification result feedback and the patient's treatment effect feedback through the feedback receiving module, and uses the lesion identification module optimization module to evaluate and update the lesion identification model, and uses the treatment suggestion optimization module to dynamically adjust and optimize the treatment suggestion.

[0011] Preferably, the image acquisition module acquires electronic film images of various tissues of patients with different ages, genders, and pathological conditions, and establishes a standardized data format and storage structure;

[0012] The image acquisition module is provided with an image upload and display interface, which supports uploading image data from multiple sources and formats, and realizes fast loading and clear display of image data under the premise of ensuring data security and patient privacy;

[0013] Among them, electronic film images include X-rays, CT images and MRI images.

[0014] Preferably, the image processing module performs image processing on the electronic film image acquired by the image acquisition module, including:

[0015] Perform image preprocessing operations including denoising, contrast enhancement and standardization on electronic film images to improve image quality and make lesions easier to identify;

[0016] Professional medical knowledge is used to annotate the pre-processed electronic film images including the location, size and type of lesions, and a historical data set is constructed.

[0017] Preferably, the lesion recognition model construction module constructs a lesion recognition model including an image feature extractor, a global average pooling layer and a second fully connected layer connected in sequence;

[0018] The image feature extractor is obtained by removing the first fully connected layer in the original convolutional neural network CNN, which is used to convolve and pool the input image;

[0019] The global average pooling layer is used to integrate the global spatial information of the entire feature map and enhance the aggregation of category features to improve classification performance;

[0020] The second fully connected layer, matched with the lesion type, is used to learn the complex relationship between features and map them to the corresponding lesion type.

[0021] Preferably, the lesion recognition model training module uses the processed electronic film image to perform transfer learning, including:

[0022] dividing the historical data set into a first historical data set and a second historical data set according to the number of electronic film images of the lesion type;

[0023] Using the first historical data set to train an image feature extractor in a lesion recognition model, and freezing parameters in the image feature extractor;

[0024] For lesion types whose number of electronic film images is greater than a preset threshold, the global average pooling layer and the second fully connected layer in the lesion recognition model are trained using the electronic film images of the lesion type in the first historical data set; for lesion types whose number of electronic film images is not greater than the preset threshold, the global average pooling layer and the second fully connected layer in the lesion recognition model are trained using the electronic film images of the lesion type in the second historical data set, and finally a trained lesion recognition model is obtained;

[0025] Among them, the number of electronic film images of each lesion type included in the first historical data set is greater than a preset threshold; the number of electronic film images of each lesion type included in the second historical data set is not greater than the preset threshold.

[0026] Preferably, the condition explanation module provides condition explanation to the patient according to the lesion identification result, including:

[0027] Use natural language processing technology to convert lesion identification results into easy-to-understand disease descriptions, and combine pictures and videos to vividly demonstrate the disease's mechanism, treatment methods, and recovery process;

[0028] At the same time, relevant health knowledge and preventive measures are recommended based on the lesion identification results and patient needs.

[0029] Preferably, the condition interaction module performs a question-and-answer type condition interaction with the patient, including:

[0030] Use natural language understanding and natural language generation technologies to interact with patients in a question-and-answer format, answer patients' questions about their conditions, and enhance their sense of participation and gain.

[0031] Preferably, the treatment suggestion generation module provides personalized treatment suggestions for the patient according to the lesion identification result, including:

[0032] Based on the results of lesion identification, the patient's medical history, genetic factors, and multi-dimensional data of living environment are analyzed, and combined with clinical guidelines and medical literature to provide patients with personalized treatment recommendations.

[0033] Preferably, the lesion identification module optimization module evaluates and updates the lesion identification model, including:

[0034] We continue to collect expert feedback on the latest electronic film images’ lesion identification results, and regularly evaluate and update the lesion identification model in combination with the latest medical research results and clinical data to ensure the accuracy of lesion identification.

[0035] Preferably, the treatment suggestion optimization module dynamically adjusts and optimizes the treatment suggestion, including:

[0036] Continuously collect patient feedback on the treatment effects of the latest treatment recommendations, dynamically adjust and optimize the treatment recommendations, continuously improve the accuracy of treatment recommendations, and achieve precision medicine.

[0037] (III) Beneficial effects

[0038] Compared with the existing technology, the personalized intelligent medical service system based on electronic film provided by the present invention realizes accurate identification of lesions in electronic film images through key technologies such as image processing, intelligent image recognition, natural language processing, system integration and deployment, continuous learning and optimization, and can provide patients with services such as condition explanation, condition interaction and personalized treatment recommendation generation, aiming to improve the quality and efficiency of medical services, optimize the allocation of medical resources, promote patient education and popular science, support clinical decision-making, promote medical research, respond to medical challenges, and meet social needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] 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.

[0040] Figure 1 It is a schematic diagram of the system of the present invention;

[0041] Figure 2 It is a schematic diagram of the process of the present invention. DETAILED DESCRIPTION

[0042] 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.

[0043] A personalized smart medical service system based on electronic film, such as Figure 1 and Figure 2 As shown, the server includes a server, which processes the electronic film image collected by the image collection module through the image processing module, and constructs a lesion recognition model using the lesion recognition model construction module. The server uses the processed electronic film image for transfer learning through the lesion recognition model training module, and uses the lesion recognition module to input the patient's electronic film image to be identified into the trained lesion recognition model to obtain a lesion recognition result;

[0044] The server provides the patient with a medical condition explanation based on the lesion identification results through the medical condition explanation module, and uses the medical condition interaction module to interact with the patient in a question-and-answer manner. The server provides the patient with personalized treatment recommendations based on the lesion identification results through the treatment recommendation generation module.

[0045] The server receives the expert's lesion identification result feedback and the patient's treatment effect feedback through the feedback receiving module, and uses the lesion identification module optimization module to evaluate and update the lesion identification model, and uses the treatment suggestion optimization module to dynamically adjust and optimize the treatment suggestions.

[0046] ① The image acquisition module collects electronic film images of various tissues of patients with different ages, genders, and pathological conditions, and establishes a standardized data format and storage structure;

[0047] The image acquisition module is equipped with an image upload and display interface, which supports uploading image data from multiple sources and formats, and realizes fast loading and clear display of image data while ensuring data security and patient privacy;

[0048] Among them, electronic film images include X-rays, CT images and MRI images.

[0049] ② The image processing module performs image processing on the electronic film image acquired by the image acquisition module, including:

[0050] Perform image preprocessing operations including denoising, contrast enhancement and standardization on electronic film images to improve image quality and make lesions easier to identify;

[0051] Professional medical knowledge is used to annotate the pre-processed electronic film images including the location, size and type of lesions, and a historical data set is constructed.

[0052] ③ The lesion recognition model construction module constructs the lesion recognition model including the image feature extractor, the global average pooling layer and the second fully connected layer connected in sequence;

[0053] The image feature extractor is obtained by removing the first fully connected layer in the original convolutional neural network CNN, which is used to convolve and pool the input image;

[0054] The global average pooling layer is used to integrate the global spatial information of the entire feature map and enhance the aggregation of category features to improve classification performance;

[0055] The second fully connected layer, matched with the lesion type, is used to learn the complex relationship between features and map them to the corresponding lesion type.

[0056] ④The lesion recognition model training module uses the processed electronic film images for transfer learning, including:

[0057] dividing the historical data set into a first historical data set and a second historical data set according to the number of electronic film images of the lesion type;

[0058] Using the first historical data set to train an image feature extractor in a lesion recognition model, and freezing parameters in the image feature extractor;

[0059] For lesion types whose number of electronic film images is greater than a preset threshold, the global average pooling layer and the second fully connected layer in the lesion recognition model are trained using the electronic film images of the lesion type in the first historical data set; for lesion types whose number of electronic film images is not greater than the preset threshold, the global average pooling layer and the second fully connected layer in the lesion recognition model are trained using the electronic film images of the lesion type in the second historical data set, and finally a trained lesion recognition model is obtained;

[0060] Among them, the number of electronic film images of each lesion type included in the first historical data set is greater than a preset threshold; the number of electronic film images of each lesion type included in the second historical data set is not greater than the preset threshold.

[0061] In the above technical solution, the lesion recognition model training module uses the processed electronic film images for transfer learning. Through transfer learning, it can effectively solve the problem of low lesion recognition accuracy caused by insufficient sample number, and ensure the recognition accuracy of the trained lesion recognition model for lesion types with insufficient sample number.

[0062] ⑤ The condition explanation module provides patients with condition explanations based on the lesion identification results, including:

[0063] Use natural language processing technology to convert lesion identification results into easy-to-understand disease descriptions, and combine pictures and videos to vividly demonstrate the disease's mechanism, treatment methods, and recovery process;

[0064] At the same time, relevant health knowledge and preventive measures are recommended based on the lesion identification results and patient needs.

[0065] ⑥ The condition interaction module conducts question-and-answer type condition interaction with the patient, including:

[0066] Use natural language understanding and natural language generation technologies to interact with patients in a question-and-answer format, answer patients' questions about their conditions, and enhance their sense of participation and gain.

[0067] ⑦ The treatment suggestion generation module provides personalized treatment suggestions for patients based on the lesion identification results, including:

[0068] Based on the results of lesion identification, the patient's medical history, genetic factors, and multi-dimensional data of living environment are analyzed, and combined with clinical guidelines and medical literature to provide patients with personalized treatment recommendations.

[0069] The above-mentioned technical solution, by providing patients with services such as condition explanation, condition interaction and generation of personalized treatment suggestions, can interact with patients in a friendly manner, better promote patient education and popular science, and help patients recover as soon as possible.

[0070] ⑧The lesion recognition module optimization module evaluates and updates the lesion recognition model, including:

[0071] We continue to collect expert feedback on the latest electronic film images’ lesion identification results, and regularly evaluate and update the lesion identification model in combination with the latest medical research results and clinical data to ensure the accuracy of lesion identification.

[0072] ⑨The treatment suggestion optimization module dynamically adjusts and optimizes the treatment suggestions, including:

[0073] Continuously collect patient feedback on the treatment effects of the latest treatment recommendations, dynamically adjust and optimize the treatment recommendations, continuously improve the accuracy of treatment recommendations, and achieve precision medicine.

[0074] The above technical solution ensures system performance and effectively improves user experience and service quality by evaluating and updating the lesion recognition model and dynamically adjusting and optimizing treatment recommendations.

[0075] 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. A personalized intelligent medical service system based on electronic film, characterized by: The server includes a server, which processes the electronic film image collected by the image collection module through the image processing module, and constructs a lesion recognition model using the lesion recognition model construction module. The server performs transfer learning using the processed electronic film image through the lesion recognition model training module, and inputs the patient's electronic film image to be recognized into the trained lesion recognition model using the lesion recognition module to obtain a lesion recognition result. The server provides the patient with a condition explanation based on the lesion identification result through the condition explanation module, and conducts a question-and-answer type condition interaction with the patient through the condition interaction module. The server provides the patient with a personalized treatment suggestion based on the lesion identification result through the treatment suggestion generation module; The server receives the expert's lesion identification result feedback and the patient's treatment effect feedback through the feedback receiving module, and uses the lesion identification module optimization module to evaluate and update the lesion identification model, and uses the treatment suggestion optimization module to dynamically adjust and optimize the treatment suggestion.

2. The personalized smart medical service system based on electronic film according to claim 1, characterized in that: The image acquisition module acquires electronic film images of various tissues of patients with different ages, genders, and pathological conditions, and establishes a standardized data format and storage structure; The image acquisition module is provided with an image upload and display interface, which supports uploading image data from multiple sources and formats, and realizes fast loading and clear display of image data under the premise of ensuring data security and patient privacy; Among them, electronic film images include X-rays, CT images and MRI images.

3. The personalized intelligent medical service system based on electronic film according to claim 2, characterized in that: The image processing module performs image processing on the electronic film image acquired by the image acquisition module, including: Perform image preprocessing operations including denoising, contrast enhancement and standardization on electronic film images to improve image quality and make lesions easier to identify; Professional medical knowledge is used to annotate the pre-processed electronic film images including the location, size and type of lesions, and a historical data set is constructed.

4. The personalized intelligent medical service system based on electronic film according to claim 1, characterized in that: The lesion recognition model construction module constructs a lesion recognition model including an image feature extractor, a global average pooling layer, and a second fully connected layer connected in sequence; The image feature extractor is obtained by removing the first fully connected layer in the original convolutional neural network CNN, which is used to convolve and pool the input image; The global average pooling layer is used to integrate the global spatial information of the entire feature map and enhance the aggregation of category features to improve classification performance; The second fully connected layer, matched with the lesion type, is used to learn the complex relationship between features and map them to the corresponding lesion type.

5. The personalized intelligent medical service system based on electronic film according to claim 4 is characterized in that: The lesion recognition model training module uses the processed electronic film image to perform transfer learning, including: dividing the historical data set into a first historical data set and a second historical data set according to the number of electronic film images of the lesion type; Using the first historical data set to train an image feature extractor in a lesion recognition model, and freezing parameters in the image feature extractor; For lesion types whose number of electronic film images is greater than a preset threshold, the global average pooling layer and the second fully connected layer in the lesion recognition model are trained using the electronic film images of the lesion type in the first historical data set; for lesion types whose number of electronic film images is not greater than the preset threshold, the global average pooling layer and the second fully connected layer in the lesion recognition model are trained using the electronic film images of the lesion type in the second historical data set, and finally a trained lesion recognition model is obtained; Among them, the number of electronic film images of each lesion type included in the first historical data set is greater than a preset threshold; the number of electronic film images of each lesion type included in the second historical data set is not greater than the preset threshold.

6. The personalized intelligent medical service system based on electronic film according to claim 1, characterized in that: The condition explanation module provides the patient with condition explanation according to the lesion identification result, including: Use natural language processing technology to convert lesion identification results into easy-to-understand disease descriptions, and combine pictures and videos to vividly demonstrate the disease's mechanism, treatment methods, and recovery process; At the same time, relevant health knowledge and preventive measures are recommended based on the lesion identification results and patient needs.

7. The personalized intelligent medical service system based on electronic film according to claim 6, characterized in that: The condition interaction module performs question-and-answer condition interaction with the patient, including: Use natural language understanding and natural language generation technologies to interact with patients in a question-and-answer format, answer patients' questions about their conditions, and enhance their sense of participation and gain.

8. The personalized smart medical service system based on electronic film according to claim 7, characterized in that: The treatment suggestion generation module provides personalized treatment suggestions to the patient according to the lesion identification results, including: Based on the results of lesion identification, the patient's medical history, genetic factors, and multi-dimensional data of living environment are analyzed, and combined with clinical guidelines and medical literature to provide patients with personalized treatment recommendations.

9. The personalized smart medical service system based on electronic film according to claim 1, characterized in that: The lesion recognition module optimization module evaluates and updates the lesion recognition model, including: We continue to collect expert feedback on the latest electronic film images’ lesion identification results, and regularly evaluate and update the lesion identification model in combination with the latest medical research results and clinical data to ensure the accuracy of lesion identification.

10. The personalized smart medical service system based on electronic film according to claim 9, characterized in that: The treatment suggestion optimization module dynamically adjusts and optimizes the treatment suggestion, including: Continuously collect patient feedback on the treatment effects of the latest treatment recommendations, dynamically adjust and optimize the treatment recommendations, continuously improve the accuracy of treatment recommendations, and achieve precision medicine.

Images