A method for processing orthopedic imaging data based on infrared images
By combining edge detection and convolutional neural network analysis of traditional orthopedic images and infrared spectral images, the accuracy problem caused by a single image processing method in orthopedic data processing is solved, and efficient and accurate diagnostic results of orthopedic image data processing are achieved.
Patent Information
- Application Number
- CN202411244541.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-06
- Publication Date
- 2024-11-05
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The prior art only relies on a single image processing method in orthopedic data processing, which makes it difficult to ensure the accuracy of diagnostic results and cannot achieve high efficiency and accuracy in orthopedic data processing.
Using orthopedic image data processing method based on infrared images, traditional orthopedic image and infrared spectral images are obtained by setting up an image processing platform, and diagnostic models are constructed using edge detection algorithms and convolutional neural networks, combining big data algorithms and Pearson correlation coefficients to carry out Analysis and joint diagnosis of image feature areas improve the accuracy of diagnostic results.
Through the joint analysis of traditional orthopedic images and infrared spectral images, the accuracy in orthopedic data processing is improved, ensuring the reliability of diagnostic results and the decision support of the medical expert team.
Smart Images

Figure CN118762010B_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the field of image processing, and in particular to an orthopedic image data processing method based on infrared images. Background Art
[0002] With the continuous development of science and technology and the development and application of digital medical imaging equipment, it has become the foundation for building digital hospitals and the guarantee for improving the level of diagnosis and medical treatment; Orthopedics in traditional Chinese medicine is a discipline that studies the prevention and treatment of injuries and diseases of the human skin, bones, qi and blood, viscera and meridians. Although the current clinical diagnostic technology is mainly based on structural imaging technologies such as X-ray, CT and MRI, with the development of science and technology, the resolution and clarity of medical infrared thermal imaging technology as a functional image are getting higher and higher. In recent years, it has been widely used in the clinical diagnosis of orthopedic diseases as one of the important auxiliary diagnostic technologies;
[0003] A data processing method based on fracture images with publication number CN108491770A discloses a data processing method based on fracture images, comprising the following steps: a) importing fracture images into a data processing center; b) performing image recognition on the imported images; c) measuring predetermined parameter values on the recognized images to obtain a set of data values; d) processing the data values obtained in step c) and outputting them. This method can help orthopedic doctors read bone images and judge the condition, greatly improving the work efficiency and quality of doctors.
[0004] The data processing method of an idiopathic scoliosis image with the publication number CN108320288A relates to a data processing method of an idiopathic scoliosis image, comprising the following steps: a) importing the idiopathic scoliosis image into a data processing center; b) performing image recognition on the imported image; c) measuring the predetermined parameter value of the recognized image to obtain a set of data values; d) processing the data values obtained in step c) and outputting them. This method can help orthopedic doctors free themselves from manual measurement of idiopathic scoliosis image processing, use data processing to perform accurate and rapid processing, and output and display the results, which can greatly improve the work efficiency and quality of surgeons.
[0005] However, at present, orthopedic data are only analyzed and processed through a single image processing method to obtain a single diagnostic result, so the accuracy of the orthopedic data processing process cannot be guaranteed. Therefore, how to improve the accuracy of the orthopedic data processing process is a problem we need to solve. To this end, a method for orthopedic image data processing based on infrared images is provided. Summary of the invention
[0006] In order to solve the above technical problems, the object of the present invention is to provide an orthopedic image data processing method based on infrared images.
[0007] The purpose of the present invention can be achieved by the following technical solution: A method for processing orthopedic image data based on infrared images, comprising the following steps:
[0008] Step S1: setting an image processing platform, setting an image data source, acquiring orthopedic image data according to the image data source, analyzing and processing the orthopedic image data, setting an image feature area, setting a diagnostic model according to the orthopedic image data, and storing the diagnostic model according to the image feature area;
[0009] Step S2: Obtain the basic medical data of the patient, and the medical staff submits a data collection application based on the basic medical data of the patient, and the image processing platform collects infrared spectral images and traditional orthopedic images based on the data collection application;
[0010] Step S3: performing edge processing on the traditional orthopedic image based on the edge detection algorithm, obtaining the image feature area in the traditional orthopedic image, analyzing and processing the image feature area according to the diagnostic model corresponding to the image feature area, determining whether there is a disease, and setting the basic lesion area according to the determination result;
[0011] Step S4: setting a correlation radius according to the basic lesion area, acquiring an infrared spectrum image, acquiring infrared data information corresponding to the pixel points corresponding to the infrared spectrum image, acquiring comprehensive infrared data information in the corresponding image feature area according to the correlation radius, judging whether there is a disease in the corresponding image feature area, and setting an infrared lesion area according to the judgment result;
[0012] Step S5: Perform a joint analysis based on the distribution of the basic lesion area and the infrared lesion area to obtain a joint diagnosis result of the corresponding image feature area.
[0013] Furthermore, the process of setting the image feature area includes:
[0014] Setting up an image processing platform, wherein an institution verification window is set up in the image processing platform, wherein the institution verification window is used to perform qualification certification on relevant medical institutions and scientific research institutions, and setting institutions that meet the qualification certification as data collection sources;
[0015] Based on the big data algorithm, the corresponding detection indicators and historical diagnosis results of the corresponding orthopedic imaging data are detected in the data collection source; the orthopedic imaging data include traditional orthopedic images and infrared spectrum images; according to the obtained traditional orthopedic images, analysis and processing are performed, and the edge data of the traditional orthopedic images are obtained based on the Canny edge detection algorithm, and the edge data are analyzed and processed to obtain the corresponding bone structure data and bone node data, wherein the bone structure data includes the position of the corresponding bone in the human body, image data and bone feature data; according to the bone structure data, an image feature area is set, and according to the bone structure data, the obtained image feature area is associated with the bone node data, and an orthopedic reference feature image is set, wherein the orthopedic reference feature image includes the distribution of human bones and the connection relationship between each other.
[0016] Furthermore, the process of setting up a diagnostic model based on orthopedic imaging data includes:
[0017] Obtain detection indicators and historical diagnosis results corresponding to orthopedic imaging data; obtain detection indicators and historical diagnosis results corresponding to the change trend of the corresponding traditional orthopedic images and the corresponding historical diagnosis results in the infrared spectrum image according to the image feature area, and set the traditional historical data set and the infrared historical data set respectively; divide the obtained traditional historical data set and infrared historical data set, and set the training set and verification set of the corresponding historical data set;
[0018] Constructing a traditional diagnostic model and an infrared diagnostic model based on a convolutional neural network; analyzing and training the traditional diagnostic model and the infrared diagnostic model according to the corresponding training set until the loss function training is stable, saving the network parameters corresponding to each model, analyzing and testing the traditional diagnostic model and the infrared diagnostic model through the corresponding test set until they meet the preset requirements, and outputting the traditional diagnostic model and the infrared diagnostic model respectively;
[0019] The corresponding diagnostic models and detection indicators are stored according to the corresponding image feature areas in the orthopedic reference feature images.
[0020] Furthermore, the process of collecting infrared spectroscopy images and traditional orthopedic images includes:
[0021] The image processing platform is provided with a patient input window, and the patient input window is used to obtain the corresponding basic medical data of the patient. The medical staff sets the corresponding data collection application form according to the basic medical data of the patient, and the data collection application form includes the corresponding basic medical data of the patient, the patient informed consent form and the medical application form; the image processing platform is provided with an infrared imaging device and a traditional imaging device, and the data collection application form is obtained, and the content in the data collection application form is verified and processed. If the verification is successful, the imaging device is granted the authority to collect images;
[0022] The image processing platform is provided with an infrared imaging device and a traditional imaging device; the infrared imaging device is used to obtain infrared spectral images; the traditional imaging device is used to obtain traditional orthopedic images, the obtained infrared spectral images and traditional orthopedic images are marked according to patient information and acquisition time, and a patient-associated data set is set according to the marking results, and the infrared spectral images and traditional orthopedic images are stored in the corresponding patient-associated data sets.
[0023] Furthermore, the process of setting the basic lesion area includes:
[0024] Obtain the corresponding traditional orthopedic image in the patient-associated data set, grayscale the corresponding traditional orthopedic image of the patient, obtain the corresponding grayscale value data in the traditional orthopedic image, and perform gradient calculation on the grayscale value data based on the Sobel gradient operator; perform analysis and processing based on the gradient calculation result, obtain the edge detection result in the traditional orthopedic image, match and analyze the obtained edge detection result with each image feature area in the orthopedic reference feature image, and obtain the corresponding image feature area based on the matching result;
[0025] Acquire the corresponding traditional diagnosis model according to the image feature area, input the obtained traditional orthopedic image into the medical diagnosis model, and output the predicted traditional diagnosis result; perform analysis and processing according to the traditional diagnosis result, and when the predicted traditional diagnosis result contains disease information, mark the image feature area as the basic lesion area;
[0026] The basic lesion area and predicted traditional diagnostic results obtained for the patient are input into a patient-associated data set, and the patient-associated data set is stored.
[0027] Furthermore, the process of obtaining comprehensive infrared data information includes:
[0028] The image processing platform obtains the corresponding patient-associated data set, and centers the corresponding basic lesion area in the patient-associated data set; presets an associated radius, wherein the associated radius is the number of bone node data separated from the corresponding basic lesion area by the center;
[0029] Obtain an orthopedic reference feature image, map the corresponding basic lesion area in the patient-related data set to the orthopedic reference feature image, and obtain an image feature area that is correlated between the orthopedic reference feature image and the basic lesion area according to the correlation radius; mark the obtained image feature area as a predicted correlation area;
[0030] An infrared spectral image is acquired, wherein the infrared spectral image includes blood data and heat data at corresponding positions of each pixel point, and corresponding data information in the infrared spectral image is mapped to corresponding pixel points in the predicted associated area, and blood supply data and heat distribution data of each pixel point in each predicted associated area are acquired, and a corresponding time series is set according to the acquisition time, and recorded as comprehensive infrared data information.
[0031] Furthermore, the process of setting the infrared lesion area includes:
[0032] Obtain the detection index corresponding to the comprehensive infrared data information and the infrared spectrum image, set the data detection standard interval according to the detection index, compare and analyze the comprehensive infrared data information corresponding to the corresponding pixel points in each predicted associated area with the corresponding data detection standard interval, if it does not belong to the corresponding data detection standard interval, the pixel point is marked as abnormal; obtain the abnormal rate of the pixel points in the predicted associated area, preset the abnormal rate threshold, when the pixel point abnormal rate is greater than or equal to the abnormal rate threshold, the predicted associated area is abnormal, when the pixel point abnormal rate is less than the abnormal rate threshold, the predicted associated area is normal;
[0033] Analyze and process the infrared data information in the predicted associated area with abnormalities, set the blood supply data change curve and the heat distribution data change curve in the corresponding predicted associated area according to the time series, input them into the infrared diagnosis model, output the predicted infrared diagnosis result, and mark the predicted associated area as an infrared lesion area;
[0034] The infrared lesion area and the predicted infrared diagnosis result obtained for the patient are input into a patient-associated data set, and the patient-associated data set is stored.
[0035] Furthermore, the process of obtaining the joint diagnosis result of the corresponding image feature area includes:
[0036] The image processing platform obtains a patient-related data set corresponding to the patient, and performs analysis and processing according to the predicted traditional diagnosis results and predicted infrared diagnosis results respectively corresponding to the basic lesion area and the infrared lesion area stored therein; and maps the basic lesion area and the infrared lesion area to the orthopedic reference feature image respectively;
[0037] According to the historical diagnosis results corresponding to the orthopedic imaging data, the correlation data of the predicted traditional diagnosis results and the predicted infrared diagnosis results are obtained based on the Pearson correlation coefficient, and a correlation threshold is preset; when the correlation data is greater than or equal to the correlation threshold, the predicted diagnosis results corresponding to the basic lesion area and the infrared lesion area are jointly analyzed to generate a joint diagnosis result; when the correlation data is less than the correlation threshold, the predicted diagnosis results corresponding to the basic lesion area and the infrared lesion area are abnormal, and the corresponding image feature area is marked as abnormal;
[0038] A preset team of medical experts will send the image feature areas marked as abnormal to the team of medical experts, who will conduct a joint diagnosis and obtain a joint diagnosis result.
[0039] Compared with the prior art, the beneficial effects of the present invention are as follows: by analyzing and processing traditional orthopedic images, corresponding image feature areas are set, and the obtained traditional orthopedic images and infrared spectral images are respectively mapped into the corresponding image feature areas, and according to whether there are abnormalities in the data of the traditional orthopedic images and the infrared spectral images in the image feature areas, basic lesion areas and infrared lesion areas are set according to the abnormal results, and a joint diagnosis is performed according to the correlation data between the corresponding disease results in each image feature area to determine whether there are abnormalities in the predicted diagnosis results or whether the corresponding patients have abnormalities, a medical expert team is set to process the data information with abnormalities, and obtain a joint diagnosis result, thereby improving the accuracy of the orthopedic data processing process to a certain extent. BRIEF DESCRIPTION OF THE DRAWINGS
[0040] Figure 1 This is a schematic diagram of an orthopedic image data processing method based on infrared images according to an embodiment of the present application. DETAILED DESCRIPTION
[0041] like Figure 1 As shown, a method for processing orthopedic image data based on infrared images comprises the following steps:
[0042] Step S1: setting an image processing platform, setting an image data source, acquiring orthopedic image data according to the image data source, analyzing and processing the orthopedic image data, setting an image feature area, setting a diagnostic model according to the orthopedic image data, and storing the diagnostic model according to the image feature area;
[0043] Step S2: Obtain the basic medical data of the patient, and the medical staff submits a data collection application based on the basic medical data of the patient, and the image processing platform collects infrared spectral images and traditional orthopedic images based on the data collection application;
[0044] Step S3: performing edge processing on the traditional orthopedic image based on the edge detection algorithm, obtaining the image feature area in the traditional orthopedic image, analyzing and processing the image feature area according to the diagnostic model corresponding to the image feature area, determining whether there is a disease, and setting the basic lesion area according to the determination result;
[0045] Step S4: setting a correlation radius according to the basic lesion area, acquiring an infrared spectrum image, acquiring infrared data information corresponding to the pixel points corresponding to the infrared spectrum image, acquiring comprehensive infrared data information in the corresponding image feature area according to the correlation radius, judging whether there is a disease in the corresponding image feature area, and setting an infrared lesion area according to the judgment result;
[0046] Step S5: Perform a joint analysis based on the distribution of the basic lesion area and the infrared lesion area to obtain a joint diagnosis result of the corresponding image feature area.
[0047] It should be further explained that, in the specific implementation process, the image processing platform is set, the image data source is set, orthopedic image data is obtained according to the image data source, the orthopedic image data is analyzed and processed, the image feature area is set, and the diagnosis model is set according to the orthopedic image data, and the diagnosis model is stored according to the image feature area. The specific implementation process includes:
[0048] An image processing platform is provided, wherein the image processing platform is used to obtain orthopedic image data, and analyze and process the obtained orthopedic image data, so as to obtain corresponding disease information;
[0049] The image processing platform is provided with an institution verification window, which is used for corresponding medical institutions, medical research institutions, etc. to enter the platform and obtain corresponding institution data information. The institution data information includes the corresponding institution name, institution qualification certificate and the corresponding certification certificate of the regulatory department. The obtained institution data information is verified and the institution corresponding to the institution data information that meets the verification is set as the data collection source; the data collection source is the data source in the platform that meets the national standard inspection, and orthopedic imaging data is collected according to the data collection source, thereby improving the accuracy of the data source;
[0050] Retrieve detection indicators and several historical diagnosis results corresponding to orthopedic imaging data based on a big data algorithm according to a data acquisition source, wherein the orthopedic imaging data includes traditional orthopedic images and infrared spectrum images; the traditional orthopedic images are CT scan images, magnetic resonance imaging images, and X-ray scan images;
[0051] The X-ray scan image is the most common image in orthopedic imaging, which shows the bone structure and plays a significant role in detecting fractures, arthritis, bone hyperplasia and other lesions;
[0052] The CT scan image is a cross-sectional image of the bone generated by X-ray imaging in multiple directions, providing high-resolution bone structure information;
[0053] The magnetic resonance imaging is a high-resolution orthopedic imaging information generated by magnetic field and radio waves, which includes the detailed structure of soft tissue and bones;
[0054] Analyze and process the obtained traditional orthopedic images, analyze and process the obtained traditional orthopedic images based on an edge detection algorithm, obtain edge data of the traditional orthopedic images based on a Canny edge detection algorithm, wherein the edge data includes bone structure data and bone node data, wherein the bone structure data includes the position of the corresponding bone in the human body, image data, and bone feature data; set an image feature area according to the bone structure data, associate the obtained image feature area according to the bone node data, and set an orthopedic reference feature image, wherein the orthopedic reference feature image includes the distribution of bones in the human body and the connection relationship between them;
[0055] Obtain detection indicators and historical diagnosis results corresponding to orthopedic imaging data; obtain detection indicators and historical diagnosis results corresponding to the change trend of the corresponding traditional orthopedic images and the corresponding historical diagnosis results in the infrared spectrum image according to the image feature area, and set the traditional historical data set and the infrared historical data set respectively; divide the obtained traditional historical data set and infrared historical data set, and set the training set and verification set of the corresponding historical data set;
[0056] Constructing a traditional diagnostic model and an infrared diagnostic model based on a convolutional neural network; analyzing and training the traditional diagnostic model and the infrared diagnostic model according to the corresponding training set until the loss function training is stable, saving the network parameters corresponding to each model, analyzing and testing the traditional diagnostic model and the infrared diagnostic model through the corresponding test set until they meet the preset requirements, and outputting the traditional diagnostic model and the infrared diagnostic model respectively;
[0057] The corresponding diagnostic models and detection indicators are stored according to the corresponding image feature areas in the orthopedic reference feature images.
[0058] It should be further explained that, in the specific implementation process, the basic medical data of the patient is obtained by the medical staff submitting a data collection application based on the basic medical data of the patient, and the image processing platform collects infrared spectral images and traditional orthopedic images based on the data collection application. The specific implementation process includes:
[0059] The image processing platform is provided with a patient input window, which is used to obtain the corresponding basic medical data of the patient, which includes the patient's name, gender, ID number, symptom information and registration information. The medical staff obtains the basic medical data of the patient according to the registration information, and the medical staff sets a corresponding data collection application according to the basic medical data of the patient, and the data collection application includes the corresponding patient information, the patient's informed consent form and the medical application form;
[0060] The image processing platform is provided with an infrared imaging device and a traditional imaging device, obtains a data collection application, verifies the contents in the data collection application, and if the verification is unsuccessful, does not grant the corresponding imaging device permission to collect orthopedic images of the corresponding patient; if the verification is successful, grants the corresponding imaging device permission to collect orthopedic images of the corresponding patient;
[0061] The infrared imaging device and the traditional imaging device complete the image acquisition of the patient corresponding to the verification data acquisition application form. Before acquiring the image of the patient, the imaging device is calibrated and corrected to ensure that the distance and angle between the device and the patient are appropriate. After calibration and correction, the infrared imaging device performs infrared spectrum analysis on the patient, obtains infrared spectrum images, and marks them according to the patient information and the corresponding acquisition time; the traditional imaging device performs orthopedic image acquisition on the patient, obtains traditional orthopedic images, and marks them according to the patient information and the corresponding acquisition time;
[0062] The traditional orthopedic images and infrared spectroscopy images collected corresponding to the patient information and their corresponding collection times are marked, a patient-related data set is set, and the obtained data information is stored in the patient-related data set.
[0063] It should be further explained that, in the specific implementation process, the edge processing of the traditional orthopedic image is performed based on the edge detection algorithm, the image feature area in the traditional orthopedic image is obtained, and the image feature area is analyzed and processed according to the diagnostic model corresponding to the image feature area to determine whether there is a disease, and the basic lesion area is set according to the judgment result. The specific implementation process includes:
[0064] The image processing platform obtains a patient-associated data set corresponding to the patient, and obtains a corresponding traditional orthopedic image in the patient-associated data set and its corresponding acquisition time;
[0065] Obtain the orthopedic reference feature images stored in the image processing platform, and analyze and process the obtained traditional orthopedic images based on the edge detection algorithm:
[0066] Grayscale the obtained traditional orthopedic image to obtain the corresponding grayscale value data in the traditional orthopedic image, perform gradient calculation on the grayscale value data based on the Sobel gradient operator, obtain the gradient amplitude and gradient direction of the corresponding pixel point in the traditional orthopedic image, obtain the local maximum value in the corresponding gradient direction of the traditional orthopedic image according to the gradient calculation result, and suppress the non-maximum points. After the suppression process is completed, the corresponding edge fragments are obtained, the edge fragments are processed based on the Hough transform algorithm, and the edge detection results in the traditional orthopedic image are obtained. Matching and analyzing the obtained edge detection results with each image feature area in the orthopedic reference feature image, and obtaining the corresponding image feature area according to the matching results;
[0067] It should be further explained that, in the specific implementation process, the process of analyzing and processing the traditional orthopedic image includes: graying the traditional orthopedic image, and the process of gradient calculation of the grayscale data includes:
[0068] ;
[0069] Where M(x,y) represents the gradient amplitude of the pixel, and θ ( x , y ) represents the gradient direction of the pixel; is the gradient value in the horizontal direction, is the gradient value in the vertical direction, where x is the horizontal coordinate of the corresponding pixel point in the traditional orthopedic image, and y is the vertical coordinate of the corresponding pixel point in the traditional orthopedic image;
[0070] Obtaining a corresponding medical diagnosis model from the matching results of the traditional orthopedic image and the image feature area in the orthopedic reference feature image, inputting the obtained traditional orthopedic image into the medical diagnosis model, and outputting a predicted traditional diagnosis result; performing analysis and processing according to the traditional diagnosis result, when the predicted traditional diagnosis result contains disease information, marking the image feature area as a basic lesion area, and when the predicted traditional diagnosis result contains no disease information, not marking the image feature area as a basic lesion area;
[0071] The basic lesion area obtained for the patient is input into a patient-associated data set, and the patient-associated data set is stored.
[0072] It should be further explained that, in the specific implementation process, the correlation radius is set according to the basic lesion area, the infrared spectrum image is obtained, the infrared data information corresponding to the pixel points corresponding to the infrared spectrum image is obtained, the comprehensive infrared data information in the corresponding image feature area is obtained according to the correlation radius, and it is judged whether there is a disease in the corresponding image feature area, and the infrared lesion area is set according to the judgment result; the infrared lesion area is judged;
[0073] The image processing platform obtains the corresponding patient-associated data set, and concentrates the corresponding basic lesion area in the patient-associated data set;
[0074] Obtain the orthopedic reference feature image stored in the image processing platform, map the corresponding basic lesion area in the patient association data set to the orthopedic reference feature image, preset an association radius, the association radius is the number of bone node data separated from the corresponding basic lesion area by the center, and obtain the image feature area within the association radius by taking the basic lesion area as the center according to the set association radius; mark the obtained image feature area as the predicted association area;
[0075] Obtaining the corresponding infrared spectrum image in the corresponding patient association data set, obtaining infrared data analysis in the predicted association area based on the infrared spectrum image, wherein the infrared spectrum image includes blood data and heat data at the corresponding position of each pixel point, mapping the corresponding data information in the infrared spectrum image to the corresponding pixel point in the predicted association area, and obtaining blood supply data and heat distribution data in each predicted association area;
[0076] The blood supply data and heat distribution data of the infrared spectrum image in the predicted associated area are sequentially obtained according to the acquisition time of the infrared spectrum image, and the time series corresponding to the blood supply data and heat distribution data are set according to the acquisition time, and recorded as comprehensive infrared data information;
[0077] Obtain the detection index corresponding to the comprehensive infrared data information and the infrared spectrum image, set the data detection standard interval according to the detection index, compare and analyze the comprehensive infrared data information corresponding to the corresponding pixel points in each predicted associated area with the corresponding data detection standard interval, if it does not belong to the corresponding data detection standard interval, the pixel point is marked as abnormal; obtain the abnormal rate of the pixel points in the predicted associated area, preset the abnormal rate threshold, when the pixel point abnormal rate is greater than or equal to the abnormal rate threshold, the predicted associated area is abnormal, when the pixel point abnormal rate is less than the abnormal rate threshold, the predicted associated area is normal;
[0078] Analyze and process the infrared data information in the predicted associated area with abnormalities, set the blood supply data change curve and the heat distribution data change curve in the corresponding predicted associated area according to the time series, input them into the infrared diagnosis model, output the predicted infrared diagnosis result, and mark the predicted associated area as an infrared lesion area;
[0079] The obtained predicted infrared diagnosis results and other data are stored in the patient-associated data set.
[0080] It should be further explained that, in the specific implementation process, the predicted diagnosis results of the basic lesion area and the infrared spectrum data are associated to obtain the joint diagnosis results, and the obtained joint diagnosis results are stored. The specific implementation process includes:
[0081] The image processing platform obtains a patient-related data set corresponding to the patient, and performs analysis and processing according to the predicted traditional diagnosis results and predicted infrared diagnosis results respectively corresponding to the basic lesion area and the infrared lesion area stored therein; and maps the basic lesion area and the infrared lesion area to the orthopedic reference feature image respectively;
[0082] According to the historical diagnosis results corresponding to the orthopedic imaging data, the correlation data of the predicted traditional diagnosis results and the predicted infrared diagnosis results are obtained based on the Pearson correlation coefficient, and a correlation threshold is preset; when the correlation data is greater than or equal to the correlation threshold, the predicted diagnosis results corresponding to the basic lesion area and the infrared lesion area are jointly analyzed to generate a joint diagnosis result; when the correlation data is less than the correlation threshold, the predicted diagnosis results corresponding to the basic lesion area and the infrared lesion area are abnormal, and the corresponding image feature area is marked as abnormal;
[0083] A preset team of medical experts will send the image feature areas marked as abnormal to the team of medical experts, who will conduct a joint diagnosis and obtain a joint diagnosis result.
[0084] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.
Claims
1. A method for processing orthopedic image data based on infrared images, characterized in that: The following steps are involved: Step S1: setting an image processing platform, setting an image data source, acquiring orthopedic image data according to the image data source, analyzing and processing the orthopedic image data, setting an image feature area, setting a diagnostic model according to the orthopedic image data, and storing the diagnostic model according to the image feature area; Step S2: Obtain the basic medical data of the patient, and the medical staff submits a data collection application based on the basic medical data of the patient, and the image processing platform collects infrared spectral images and traditional orthopedic images based on the data collection application; Step S3: performing edge processing on the traditional orthopedic image based on the edge detection algorithm, obtaining the image feature area in the traditional orthopedic image, analyzing and processing the image feature area according to the diagnostic model corresponding to the image feature area, determining whether there is a disease, and setting the basic lesion area according to the determination result; Step S4: setting a correlation radius according to the basic lesion area, acquiring an infrared spectrum image, acquiring infrared data information corresponding to the pixel points corresponding to the infrared spectrum image, acquiring comprehensive infrared data information in the corresponding image feature area according to the correlation radius, judging whether there is a disease in the corresponding image feature area, and setting an infrared lesion area according to the judgment result; Step S5: performing a joint analysis based on the distribution of the basic lesion area and the infrared lesion area to obtain a joint diagnosis result of the corresponding image feature area; The process of obtaining comprehensive infrared data information includes: The image processing platform obtains the corresponding patient-associated data set, and centers the corresponding basic lesion area in the patient-associated data set; presets an associated radius, wherein the associated radius is the number of bone node data separated from the corresponding basic lesion area by the center; Obtain an orthopedic reference feature image, map the corresponding basic lesion area in the patient-related data set to the orthopedic reference feature image, and obtain an image feature area that is correlated between the orthopedic reference feature image and the basic lesion area according to the correlation radius; mark the obtained image feature area as a predicted correlation area; Acquire an infrared spectrum image, wherein the infrared spectrum image includes blood data and heat data at corresponding positions of each pixel point, map corresponding data information in the infrared spectrum image to corresponding pixel points in the predicted associated area, acquire blood supply data and heat distribution data of each pixel point in each predicted associated area, set a corresponding time series according to the acquisition time, and record it as comprehensive infrared data information; The process of setting the infrared lesion area includes: Obtain the detection index corresponding to the comprehensive infrared data information and the infrared spectrum image, set the data detection standard interval according to the detection index, compare and analyze the comprehensive infrared data information corresponding to the corresponding pixel points in each predicted associated area with the corresponding data detection standard interval, and if it does not belong to the corresponding data detection standard interval, mark the pixel point as abnormal; obtain the abnormal rate of the pixel points in the predicted associated area, preset the abnormal rate threshold, when the pixel point abnormal rate is greater than or equal to the abnormal rate threshold, the predicted associated area is abnormal, and when the pixel point abnormal rate is less than the abnormal rate threshold, the predicted associated area is normal; Analyze and process the infrared data information in the predicted associated area with abnormalities, set the blood supply data change curve and the heat distribution data change curve in the corresponding predicted associated area according to the time series, input them into the infrared diagnosis model, output the predicted infrared diagnosis result, and mark the predicted associated area as an infrared lesion area; The infrared lesion area and the predicted infrared diagnosis result obtained for the patient are input into a patient-associated data set, and the patient-associated data set is stored.
2. The orthopedic image data processing method based on infrared images according to claim 1 is characterized in that: The process of setting the image feature area includes: Setting up an image processing platform, wherein an institution verification window is set up in the image processing platform, wherein the institution verification window is used to perform qualification certification on relevant medical institutions and scientific research institutions, and setting institutions that meet the qualification certification as data collection sources; Based on the big data algorithm, the corresponding detection indicators and historical diagnosis results of the corresponding orthopedic imaging data are detected in the data collection source; the orthopedic imaging data include traditional orthopedic images and infrared spectrum images; according to the obtained traditional orthopedic images, analysis and processing are performed, and the edge data of the traditional orthopedic images are obtained based on the Canny edge detection algorithm, and the edge data are analyzed and processed to obtain the corresponding bone structure data and bone node data, wherein the bone structure data includes the position of the corresponding bone in the human body, image data and bone feature data; according to the bone structure data, an image feature area is set, and according to the bone structure data, the obtained image feature area is associated with the bone node data, and an orthopedic reference feature image is set, wherein the orthopedic reference feature image includes the distribution of human bones and the connection relationship between each other.
3. The orthopedic image data processing method based on infrared images according to claim 2 is characterized in that: The process of setting up a diagnostic model based on orthopedic imaging data includes: Obtain detection indicators and historical diagnosis results corresponding to orthopedic imaging data; obtain detection indicators and historical diagnosis results corresponding to the change trend of the corresponding traditional orthopedic images and the corresponding historical diagnosis results in the infrared spectrum image according to the image feature area, and set the traditional historical data set and the infrared historical data set respectively; divide the obtained traditional historical data set and infrared historical data set, and set the training set and verification set of the corresponding historical data set; Constructing a traditional diagnostic model and an infrared diagnostic model based on a convolutional neural network; analyzing and training the traditional diagnostic model and the infrared diagnostic model according to the corresponding training set until the loss function training is stable, saving the network parameters corresponding to each model, analyzing and testing the traditional diagnostic model and the infrared diagnostic model through the corresponding test set until they meet the preset requirements, and outputting the traditional diagnostic model and the infrared diagnostic model respectively; The corresponding diagnostic models and detection indicators are stored according to the corresponding image feature areas in the orthopedic reference feature images.
4. The orthopedic image data processing method based on infrared images according to claim 3 is characterized in that: The process of collecting infrared spectrum images and traditional orthopedic images includes: The image processing platform is provided with a patient input window, and the patient input window is used to obtain the corresponding basic medical data of the patient. The medical staff sets the corresponding data collection application form according to the basic medical data of the patient, and the data collection application form includes the corresponding basic medical data of the patient, the patient informed consent form and the medical application form; the image processing platform is provided with an infrared imaging device and a traditional imaging device, and the data collection application form is obtained, and the content in the data collection application form is verified and processed. If the verification is successful, the imaging device is granted the authority to collect images; The image processing platform is provided with an infrared imaging device and a traditional imaging device; the infrared imaging device is used to obtain infrared spectral images; the traditional imaging device is used to obtain traditional orthopedic images, the obtained infrared spectral images and traditional orthopedic images are marked according to patient information and acquisition time, and a patient-associated data set is set according to the marking results, and the infrared spectral images and traditional orthopedic images are stored in the corresponding patient-associated data sets.
5. The orthopedic image data processing method based on infrared images according to claim 4 is characterized in that: The process of setting the basic lesion area includes: Obtain the corresponding traditional orthopedic image in the patient-associated data set, grayscale the corresponding traditional orthopedic image of the patient, obtain the corresponding grayscale value data in the traditional orthopedic image, and perform gradient calculation on the grayscale value data based on the Sobel gradient operator; perform analysis and processing based on the gradient calculation result, obtain the edge detection result in the traditional orthopedic image, match and analyze the obtained edge detection result with each image feature area in the orthopedic reference feature image, and obtain the corresponding image feature area based on the matching result; Acquire the corresponding traditional diagnosis model according to the image feature area, input the obtained traditional orthopedic image into the medical diagnosis model, and output the predicted traditional diagnosis result; perform analysis and processing according to the traditional diagnosis result, and when the predicted traditional diagnosis result contains disease information, mark the image feature area as the basic lesion area; The basic lesion area and predicted traditional diagnostic results obtained for the patient are input into a patient-associated data set, and the patient-associated data set is stored.
6. The orthopedic image data processing method based on infrared images according to claim 5 is characterized in that: The process of obtaining the joint diagnosis result of the corresponding image feature area includes: The image processing platform obtains a patient-related data set corresponding to the patient, and performs analysis and processing according to the predicted traditional diagnosis results and predicted infrared diagnosis results respectively corresponding to the basic lesion area and the infrared lesion area stored therein; and maps the basic lesion area and the infrared lesion area to the orthopedic reference feature image respectively; According to the historical diagnosis results corresponding to the orthopedic imaging data, the correlation data of the predicted traditional diagnosis results and the predicted infrared diagnosis results are obtained based on the Pearson correlation coefficient, and a correlation threshold is preset; when the correlation data is greater than or equal to the correlation threshold, the predicted diagnosis results corresponding to the basic lesion area and the infrared lesion area are jointly analyzed to generate a joint diagnosis result; when the correlation data is less than the correlation threshold, the predicted diagnosis results corresponding to the basic lesion area and the infrared lesion area are abnormal, and the corresponding image feature area is marked as abnormal; A preset team of medical experts will send the image feature areas marked as abnormal to the team of medical experts, who will conduct a joint diagnosis and obtain a joint diagnosis result.
