Orthopedic image data analysis platform and method

Abstract

The invention discloses an orthopaedic image data analysis platform and method, and relates to the technical field of orthopaedic images, and the platform comprises a data collection module which obtains related information of a patient from a plurality of data sources, and comprises an image data unit which obtains orthopaedic related image data and 3D imaging data of a corresponding part of the patient through equipment; the data processing module is used for processing and fusing the multi-modal data acquired by the data acquisition module, and comprises an image data processing unit for performing image enhancement, denoising and registration on the image data; the modeling unit is used for converting the 3D imaging data into a corresponding three-dimensional anatomical model, docking the image data with the three-dimensional anatomical model and mapping the image data to the three-dimensional anatomical model to obtain a preoperative model; the navigation module performs angle matching and dynamic correction by using a preoperative model and real-time image data, helps a doctor to monitor the position and anatomical structure change of a patient in real time during an operation, and can improve the accuracy of the operation.

Description

Technical Field

[0001] The present invention relates to the field of orthopedic imaging technology, and in particular to an orthopedic imaging data analysis platform and method. Background Art

[0002] With the rapid development of modern medical imaging technology, imaging data plays an increasingly important role in the diagnosis, treatment and prognosis of orthopedic diseases. Common types of orthopedic images include X-rays, CT scans, MRIs and ultrasound images. These imaging data provide valuable information for clinicians, helping them make more accurate diagnosis and treatment decisions in clinical practice. With the rapid development of computer vision and artificial intelligence technology, automated analysis methods based on image recognition and deep learning have gradually emerged and have made certain progress in the field of medical image analysis.

[0003] After searching, the Chinese patent with publication number CN118490364A discloses a full orthopedic platform surgical robot and its navigation method. The patent includes a core control system that is an embedded high-performance microprocessor, which is used to coordinate the workflow between various components and support multi-task concurrent processing; a modular system installed at the end of the surgical robot; a high-precision positioning system for real-time tracking of the position of surgical instruments in space; and an intelligent image fusion system for fusing preoperative images with real-time images of the surgical site, and planning and confirming the surgical path based on the fused image data.

[0004] In the prior art, orthopedic imaging data analysis platforms mainly focus on the processing of static images and the preliminary diagnosis of lesions, lacking the ability of personalization and dynamic tracking. The processing of imaging data is mostly concentrated in the preoperative diagnosis stage. During the patient treatment process, it fails to timely reflect the changes in the injured part through real-time imaging data updating and monitoring, and lacks the analysis capability combined with patient recovery data and big data resources. Therefore, the present invention proposes an orthopedic imaging data analysis platform and method. Summary of the invention

[0005] The purpose of the present invention is to provide an orthopedic imaging data analysis platform and method to solve the problems mentioned in the above background technology.

[0006] An orthopedic imaging data analysis platform, comprising:

[0007] A data acquisition module, which uses multiple data units to acquire the original data of the patient, wherein the data acquisition module includes an image data unit, a clinical data unit and a sensor data unit;

[0008] The data processing module processes and fuses the raw data collected by the data acquisition module using the image data processing unit, the modeling unit and the data monitoring unit to obtain a preoperative model;

[0009] The navigation module integrates the preoperative model as the basic navigation data for the operation. The navigation module also obtains the patient's position information in real time during the operation, compares it with the preoperative model, and performs dynamic angle matching and correction.

[0010] Preferably, in the data acquisition module:

[0011] An image data unit, which acquires orthopedic-related image data and 3D imaging data of the patient's corresponding injured part through relevant equipment, and respectively configures each image data and each 3D imaging data into an image data group and a 3D imaging data group;

[0012] Clinical data unit, used to record the patient's medical history, physical examination data, and laboratory test data;

[0013] The sensor data unit obtains the patient's physiological data and self-reported data through wearable devices or sensors.

[0014] More preferably, the related equipment includes CT equipment, X-ray equipment, and MRI equipment;

[0015] The physiological data include the patient's exercise volume, joint mobility, gait, and range of motion; the self-reported data include pain scores and symptom changes.

[0016] Preferably, in the data processing module, the image data processing unit uses medical image processing technology to perform image enhancement, denoising and registration on the image data;

[0017] Specifically, the image data processing unit configures the image data first acquired by the image data unit according to the arrangement of the time series composed of the time points of the date changes as the reference image data, and the image data in the image data group other than the reference image data is configured as the subsequent image data. After the image data processing unit processes the reference image data using the medical image processing technology, it extracts the key bone features therein, and calculates the angle information of the reference image data according to the key bone features;

[0018] The image data processing unit uses the registration mechanism to calculate the angular deviation of the coordinates of the key bone points of the key bone features between the subsequent image data and the baseline image data, and obtains the angular difference between the subsequent image data and the baseline image data by measuring the transformation parameters such as rotation and translation of the subsequent image data after registration compared with the baseline image data.

[0019] Furthermore, a matching angle threshold is preset inside the image data processing unit to filter out subsequent image data with a large angle difference. After the image data processing unit performs angle matching on the subsequent image data and the reference image data, when the angle difference is less than the set matching angle threshold, it is considered to be qualified image data, and the qualified subsequent image data is arranged and displayed along a time series.

[0020] Furthermore, the image data processing unit also presets an associated angle threshold value, which is used to determine subsequent image data that has similarity with the target image data although the angle difference does not match the matching angle threshold value, wherein the target image data includes the reference image data, or the subsequent image data at a date and time point adjacent to the reference image data;

[0021] After the image data processing unit performs angle matching on the subsequent image data and the reference image data, when the angle difference of the subsequent image data exceeds the matching angle threshold, if a group of image data has similarity with a certain image data in the corresponding time series and meets the associated angle threshold requirement, the image data is assigned to a time point in the corresponding time series, and an expandable copy is created at the time point to store the image data matching the associated angle threshold;

[0022] The similarity is determined by the following formula:

[0023]

[0024] Where, I, J represent the two image data to be compared; μ I ,μ J are the average brightness of images I and J, respectively, and σ J are the brightness variances of images I and J, respectively, representing the contrast information of the images, and σ IJ is the covariance of images I and J, which is used to measure the structural similarity between the two images, and C1 and C2 are constants used to stabilize the calculation and avoid the denominator being zero; θ is the angle difference between the current image data and the reference image data; θ m is the matching angle threshold, θ a is the association angle threshold, α is a constant that controls the intensity of the influence of angle deviation and similarity;

[0025] When θ≤θ m When , the structural similarity of the two image data is only determined by their content and has nothing to do with the angle;

[0026] When θ m <θ<θ aWhen , the structural similarity will be gradually affected by the angle difference, and the image similarity value will gradually decrease, reflecting the influence of angle deviation;

[0027] When θ is greater than θ a When , the similarity is set to 0, indicating that the angle difference is too large and the image data no longer meets the matching or association conditions.

[0028] Preferably, in the data processing module, the modeling unit converts the 3D imaging data group acquired by the data acquisition module into a three-dimensional anatomical model corresponding to the patient's torso, and the modeling unit connects the image data corresponding to the last time point arranged in time series in the image data group processed by the image data processing unit with the three-dimensional anatomical model, maps the image data to the three-dimensional anatomical model, and obtains a preoperative model.

[0029] Preferably, in the data processing module, the data monitoring unit arranges each image data in the time series by time, detects the injured part of the patient, and matches it with the big data to obtain the recovery data.

[0030] Furthermore, the working method of the data monitoring unit comprises the following steps:

[0031] S1, arranging the image data processed by the image data processing unit according to time, extracting the regional features of the injured part, and calculating the change range of the image data at each time point;

[0032] S2, extracting data based on the characteristics of each time point, drawing a dynamic recovery curve to show the recovery progress;

[0033] S3, extract identification information based on the regional characteristics of the patient's injured part, and use a similarity matching algorithm to match it with similar cases in the orthopedic database;

[0034] S4, extracting corresponding restoration data benchmarks from the orthopedic database based on matched similar cases;

[0035] S5, comparing the patient's actual recovery curve with the recovery benchmark of the big data, calculating the recovery deviation and obtaining the recovery data;

[0036] S6. Generate a personalized recovery plan based on the comparison results.

[0037] The present invention also provides an orthopedic imaging data analysis method, using an orthopedic imaging data analysis platform according to any one of claims 1 to 9, the method comprising the following steps:

[0038] Step 1: Acquire the patient's original data, including image data and 3D imaging data, through multiple data units of the data acquisition module;

[0039] Step 2: The data processing module converts the acquired 3D imaging data into a 3D anatomical model corresponding to the patient's torso and sorts each image data in time series;

[0040] Step 3: The data processing module extracts regional features from the image data in the time series and generates a dynamic recovery curve;

[0041] And the data processing module matches the dynamic recovery curve with the orthopedic database to assist doctors in judging whether the patient's recovery meets expectations;

[0042] Step 4: During the operation, the data processing module connects the image data at the last time point in the time series with the three-dimensional anatomical model, maps the image data to the three-dimensional anatomical model, and obtains the preoperative model;

[0043] At the same time, the navigation module integrates the preoperative model and obtains the patient's position information in real time during the operation, compares it with the preoperative model, and performs dynamic angle matching and correction.

[0044] The beneficial effects of the present invention are:

[0045] The navigation module in the present invention uses preoperative models and real-time image data for angle matching and dynamic correction, helping doctors to monitor the patient's injury position and anatomical structure changes in real time during the operation, which can improve the accuracy of the operation and reduce the risk of misoperation; and through the fusion of image data, clinical data and sensor data from the data acquisition module, the solution can integrate the patient's medical history, physical examination, movement status and other information to provide a comprehensive recovery status assessment, which can better reflect the patient's health status.

[0046] The time series dynamic recovery curve in the present invention shows the recovery progress of the injured part and quantifies various recovery indicators such as bone density, fracture healing, soft tissue repair, clinical health score and sports score. The dynamic recovery curve can not only intuitively show the recovery progress, but also be compared with big data benchmarks to help doctors understand the patient's recovery status in a timely manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] 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 description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0048] Figure 1 The figure is a flow chart of an orthopedic imaging data analysis method in the present invention. DETAILED DESCRIPTION

[0049] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only 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.

[0050] See also Figure 1 As shown, the present invention provides an orthopedic imaging data analysis platform, including a data acquisition module, a data processing module and a navigation module.

[0051] The data acquisition module uses multiple data units to obtain the patient's original data, including: image data unit, clinical data unit and sensor data unit.

[0052] The image data unit obtains orthopedic-related image data and 3D imaging data of the patient's corresponding injured part through relevant equipment in a time series arrangement of date alternation, for example, "year-month-day". For example, the present application obtains orthopedic-related image data and 3D imaging data of the patient's corresponding injured part once a day within a XX time period through relevant equipment according to date alternation; and configures each image data and each 3D imaging data into an image data group and a 3D imaging data group respectively. In this embodiment, the relevant equipment includes CT equipment, X-ray equipment, and MRI equipment.

[0053] Clinical data unit, used to record the patient's medical history, physical examination data, and laboratory test data.

[0054] The sensor data unit obtains the patient's physiological data and self-reported data through the patient's self-report and wearable devices or sensors. The physiological data includes the patient's exercise volume, joint mobility, gait, and range of motion; the self-reported data includes pain scores and symptom changes.

[0055] It should be noted that the self-report is the WOMAC scale for joint pain, while the pain score is a subjective quantitative indicator obtained through self-report by patients; symptom changes can be obtained through self-report and wearable devices or sensors. Wearable devices or sensors are accelerometers and gyroscopes.

[0056] In this technical solution, the data processing module uses the image data processing unit and the modeling unit to process and fuse the original data collected by the data acquisition module.

[0057] Among them, the image data processing unit uses medical image processing technology to perform image enhancement, denoising and registration on each image data in the image data group; specifically, the image data processing unit configures the image data first acquired by the image data unit according to the arrangement of the date replacement time series as the reference image data, and after processing the reference image data using the image enhancement and denoising technology disclosed in the patent with publication number CN118490364A, extracts the key bone features therein, such as joints and bone edges, and calculates the angle information of the reference image data based on the key bone features.

[0058] It should be noted that the calculation method of the angle information of the reference image data includes:

[0059] Step 1: Identify the bone contour of the key bone feature through the Canny edge detection method.

[0060] Step 2: Use key point detection algorithms, such as Harris corner points and SIFT feature points, to determine the coordinates of the key bone points of the key bone features.

[0061] Step 3: Calculate the obtained coordinates of the skeleton key points through the vector angle formula to obtain the angle information of the reference image data.

[0062] By calculating and acquiring the angle information of the reference image data as above, the reference image data can be used as an angle reference for other image data in the time series in terms of angle information.

[0063] Specifically, when using the image data processing unit to process subsequent image data in the image data group except the baseline image data, a registration mechanism is used to calculate the angular deviation of the coordinates of the skeletal key points of the key skeletal features between the subsequent image data and the baseline image data, and by measuring the rotation, translation and other transformation parameters of the subsequent image data after registration compared with the baseline image data, the angular difference between the subsequent image data and the baseline image data is obtained.

[0064] In the above embodiment, the angle deviation is the angle error caused by the rotation or change between the subsequent image data and the reference image data, which is calculated by transforming parameters such as the difference in rotation angle; the angle difference is the angle difference observed between the actual image data, which is the result of the angle deviation and is used for subsequent matching and filtering judgment;

[0065] In addition, a matching angle threshold is preset inside the image data processing unit. The matching angle threshold is used to filter out subsequent image data with large angle differences, which means that when the angle difference between the subsequent image data and the baseline image data is within this threshold range, they are considered to meet the matching conditions, and the subsequent image data that meet the matching conditions are arranged and displayed along the time axis in the image data processing unit; and when each subsequent image data is arranged and displayed along the time axis, the corresponding shooting date, time and key information of the patient are displayed on each subsequent image data, which is convenient for medical staff to observe and track dynamically.

[0066] On this basis, an associated angle threshold is also set in the image data processing unit. The associated angle threshold is used to determine subsequent image data whose angle differences do not match the matching angle threshold but still have similarity or association with the target image data, wherein the target image data includes: baseline image data, or subsequent image data at adjacent time points to the baseline image data.

[0067] Relevance refers to the similarity between image data and the reference image in terms of spatial structure, time point or morphological characteristics, which is quantified by similarity metrics or angle deviation and is used to detect image data that have similar structures or overlap in temporal characteristics with the reference image.

[0068] Specifically, when the angle difference of subsequent image data exceeds the matching angle threshold range, if a group of image data has a similarity with a certain image data in the corresponding time series that meets the association threshold requirement, the image data is assigned to the corresponding time point in the time axis, and an expandable copy is created at that time point to accommodate the image data that can match the corresponding association angle threshold.

[0069] The similarity refers to the similarity between the subsequent image data and the target image data in the following features, including:

[0070] Structural features: whether the bone contours, joint morphology, and soft tissue morphology in the subsequent image data are similar to the target image data;

[0071] Angle feature: whether the difference between the angle deviation of the subsequent image data and the angle deviation of the target image angle is within an acceptable range;

[0072] Texture and grayscale distribution: whether the grayscale distribution and texture features of subsequent image data are close to the target image data.

[0073] Based on the above characteristics, the similarity is determined by the following formula:

[0074]

[0075] Where, I, J represent the two image data to be compared; μ I ,μ J are the average brightness of images I and J, respectively, and σ J are the brightness variances of images I and J, respectively, representing the contrast information of the images, and σ IJ is the covariance of images I and J, which is used to measure the structural similarity between the two images, and C1 and C2 are constants used to stabilize the calculation and avoid the denominator being zero; θ is the angle difference between the current image data and the reference image data; θ m is the matching angle threshold, θ a is the association angle threshold, α is a constant that controls the intensity of the influence of angle deviation and similarity;

[0076] When θ≤θ m When , the structural similarity of the two image data is only determined by their content and has nothing to do with the angle deviation;

[0077] When θ m <θ<θ a When , the structural similarity will be gradually affected by the angle difference, and the image similarity value will gradually decrease, reflecting the influence of angle deviation;

[0078] When θ is greater than θ a When , the similarity is set to 0, indicating that the angle difference is too large and the image data no longer meets the matching or association conditions;

[0079] The modeling unit converts the 3D imaging data collected by the image data unit into a three-dimensional anatomical model corresponding to the patient's torso, and locates the part of interest for orthopedic surgery in the model. The modeling unit also connects the image data processed by the image data processing unit with the three-dimensional anatomical model, maps the image data to the three-dimensional anatomical model, and obtains a preoperative model;

[0080] Mapping includes image registration technology, image reconstruction and 3D modeling technology, and image data and model fusion technology;

[0081] Image registration is a technology that aligns image data from different sources, different times or different modalities in the same space to ensure that the spatial coordinates of each data are consistent;

[0082] Image reconstruction and 3D modeling generate three-dimensional anatomical models from the acquired CT or CT by surface reconstruction method;

[0083] The fusion technology of image data and model maps the registered and reconstructed image data with the three-dimensional model to form a preoperative model;

[0084] The modeling unit performs multi-channel processing on the constructed three-dimensional anatomical model to extract key information, including the location of bone damage, joint surface structure, and the relative positions of blood vessels and nerves;

[0085] Among them, multi-channel processing refers to the use of CT, MRI, and X-ray in medical image processing to perform multi-scale and multi-feature analysis on single image data to extract different types of information;

[0086] The surgical site is determined based on bone density analysis and edge detection to determine the location of bone damage. Bone density analysis is based on CT images, and edge detection is based on X-rays.

[0087] Articular surface structure: The MRI image is segmented into articular surface and soft tissue by threshold segmentation method, and the range of motion of the joint is determined by angle calculation, providing reference information for joint replacement;

[0088] The relative positions of blood vessels and nerves: identifying blood vessels and nerves through MRI images and generating corresponding spatial position markers to avoid accidental injuries during surgery;

[0089] The data monitoring unit detects the injured part of the patient by arranging the various image data in the timeline according to time, and matches them with the big data to obtain the recovery data;

[0090] The working method of the data monitoring unit comprises the following steps:

[0091] S1. Arrange the image data processed by the image data processing unit according to the time axis, extract the regional features of the injured part, including the fracture length, bone density change, soft tissue injury degree, joint mobility, and calculate the change amplitude of the image data at each time point to quantify the recovery progress of the injured part;

[0092] S2. Draw a dynamic recovery curve based on the regional characteristics at each time point to show the recovery progress;

[0093] The dynamic recovery curve adopts the formula:

[0094] R(t)=ωB×B(t)+ωF×F(t)+ωT×T(t)+ωC×C(t)+ωS×S(t);

[0095] Where R(t) is the patient’s recovery status value at time t;

[0096] B(t) is the change of bone density over time, which comes from the image data processing unit, and is analyzed by the gray value of the CT image;

[0097] F(t) is the change of fracture healing status over time, which comes from the image data processing unit and is estimated by image edge detection and the change of fracture length;

[0098] T(t) is the change of soft tissue repair over time, which comes from the image data processing unit and is evaluated by MRI images;

[0099] C(t) is the health status score from the clinical data unit, which is obtained based on the patient's medical history, physical examination data, and laboratory test data;

[0100] S(t) is the motion score from the sensor data unit, which is obtained based on the patient's physiological data and self-report data;

[0101] ωB, ωF, ωT, ωC, ωS are weight coefficients of corresponding items, and ωB+ωF+ωT+ωC+ωS=1;

[0102] S3. Based on the regional characteristics of the patient's injured part, extract identification information, including injury type, soft tissue injury type, and recovery progress, and use a similarity matching algorithm to match with similar cases in the orthopedic database. Specifically, use the cosine similarity algorithm for matching;

[0103] S4. Based on the matched similar cases, the corresponding recovery data benchmarks are extracted from the orthopedic database, including healing time, recovery curve, surgery and rehabilitation recommendations;

[0104] S5. Compare the patient's actual recovery curve with the recovery benchmark of big data and calculate the recovery deviation. Deviation analysis can help doctors determine whether the patient's recovery meets expectations.

[0105] Specifically, the actual recovery curve is generated by summarizing the multimodal data of connected time points during the patient's treatment process, including data collection, feature extraction and quantification of CT images, X-rays, and MRI images, and dynamic recovery curve generation and visualization;

[0106] The curve generation includes a horizontal axis and a vertical axis, the horizontal axis is time, and the vertical axis is the patient's recovery status value;

[0107] Generate a personalized recovery plan based on the comparison results;

[0108] The navigation module integrates the preoperative model as the basic navigation data for the operation, and identifies the key points, landmarks, fracture lines and joint surfaces related to the patient's anatomical structure. During the operation, the navigation module obtains the patient's position information in real time, compares it with the preoperative model, and performs dynamic angle matching and correction.

[0109] The modeling unit uses the image data of the last time point in the timeline to connect with the three-dimensional anatomical model.

[0110] Finally, it should be noted that the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An orthopedic imaging data analysis platform, characterized in that: include: A data acquisition module, which uses multiple data units to acquire the original data of the patient, wherein the data acquisition module includes an image data unit, a clinical data unit and a sensor data unit; The data processing module processes and fuses the raw data collected by the data acquisition module using the image data processing unit, the modeling unit and the data monitoring unit to obtain a preoperative model; The navigation module integrates the preoperative model as the basic navigation data for the operation. The navigation module also obtains the patient's position information in real time during the operation, compares it with the preoperative model, and performs dynamic angle matching and correction.

2. The orthopedic imaging data analysis platform according to claim 1, characterized in that: In the data acquisition module: An image data unit, which acquires orthopedic-related image data and 3D imaging data of the patient's corresponding injured part through relevant equipment, and respectively configures each image data and each 3D imaging data into an image data group and a 3D imaging data group; Clinical data unit, used to record the patient's medical history, physical examination data, and laboratory test data; The sensor data unit obtains the patient's physiological data and self-reported data through wearable devices or sensors.

3. The orthopedic imaging data analysis platform according to claim 2, characterized in that: The related equipment includes CT equipment, X-ray equipment, and MRI equipment; The physiological data include the patient's exercise volume, joint mobility, gait, and range of motion; the self-reported data include pain scores and symptom changes.

4. The orthopedic imaging data analysis platform according to claim 1, characterized in that: In the data processing module, the image data processing unit uses medical image processing technology to perform image enhancement, denoising and registration on the image data; Specifically, the image data processing unit configures the image data first acquired by the image data unit according to the arrangement of the time series composed of the time points of the date changes as the reference image data, and the image data in the image data group other than the reference image data is configured as the subsequent image data. After the image data processing unit processes the reference image data using the medical image processing technology, it extracts the key bone features therein, and calculates the angle information of the reference image data according to the key bone features; The image data processing unit uses the registration mechanism to calculate the angular deviation of the coordinates of the key bone points of the key bone features between the subsequent image data and the baseline image data, and obtains the angular difference between the subsequent image data and the baseline image data by measuring the transformation parameters such as rotation and translation of the subsequent image data after registration compared with the baseline image data.

5. The orthopedic imaging data analysis platform according to claim 4, characterized in that: A matching angle threshold is also preset inside the image data processing unit to filter out subsequent image data with a large angle difference. After the image data processing unit performs angle matching on the subsequent image data and the reference image data, when the angle difference is less than the set matching angle threshold, it is considered to be qualified image data, and the qualified subsequent image data is arranged and displayed along a time series.

6. The orthopedic imaging data analysis platform according to claim 5, characterized in that: The image data processing unit also presets an associated angle threshold inside, which is used to determine subsequent image data that has similarity with the target image data although the angle difference does not match the matching angle threshold, wherein the target image data includes the reference image data, or the subsequent image data at a date and time point adjacent to the reference image data; After the image data processing unit performs angle matching on the subsequent image data and the reference image data, when the angle difference of the subsequent image data exceeds the matching angle threshold, if a group of image data has similarity with a certain image data in the corresponding time series and meets the associated angle threshold requirement, the image data is assigned to a time point in the corresponding time series, and an expandable copy is created at the time point to store the image data matching the associated angle threshold; The similarity is determined by the following formula: Where, I, J represent the two image data to be compared; μ I ,μ J are the average brightness of images I and J, respectively, and σ I ,σ J are the brightness variances of images I and J, respectively, representing the contrast information of the images, and σ IJ is the covariance of images I and J, which is used to measure the structural similarity between the two images, and C1 and C2 are constants used to stabilize the calculation and avoid the denominator being zero; θ is the angle difference between the current image data and the reference image data; θ m is the matching angle threshold, θ a is the association angle threshold, α is a constant that controls the intensity of the influence of angle deviation and similarity; When θ≤θ m When , the structural similarity of the two image data is only determined by their content and has nothing to do with the angle; When θ m <θ<θ a When , the structural similarity will be gradually affected by the angle difference, and the image similarity value will gradually decrease, reflecting the influence of angle deviation; When θ is greater than θ a When , the similarity is set to 0, indicating that the angle difference is too large and the image data no longer meets the matching or association conditions.

7. The orthopedic imaging data analysis platform according to claim 6, characterized in that: In the data processing module, the modeling unit converts the 3D imaging data group acquired by the data acquisition module into a three-dimensional anatomical model corresponding to the patient's torso, and the modeling unit connects the image data corresponding to the last time point arranged in time series in the image data group processed by the image data processing unit with the three-dimensional anatomical model, maps the image data to the three-dimensional anatomical model, and obtains a preoperative model.

8. The orthopedic imaging data analysis platform according to claim 7, characterized in that: In the data processing module, the data monitoring unit arranges each image data in the time series according to time, detects the injured part of the patient, and matches it with the big data to obtain the recovery data.

9. The orthopedic imaging data analysis platform according to claim 8, characterized in that: The working method of the data monitoring unit comprises the following steps: S1, arranging the image data processed by the image data processing unit according to time, extracting the regional features of the injured part, and calculating the change range of the image data at each time point; S2, extracting data based on the characteristics of each time point, drawing a dynamic recovery curve to show the recovery progress; S3, extract identification information based on the regional characteristics of the patient's injured part, and use a similarity matching algorithm to match it with similar cases in the orthopedic database; S4, extracting corresponding restoration data benchmarks from the orthopedic database based on matched similar cases; S5, comparing the patient's actual recovery curve with the recovery benchmark of the big data, calculating the recovery deviation and obtaining the recovery data; S6. Generate a personalized recovery plan based on the comparison results.

10. An orthopedic image data analysis method, using an orthopedic image data analysis platform according to any one of claims 1 to 9, characterized in that: The method comprises the following steps: Step 1: Acquire the patient's original data, including image data and 3D imaging data, through multiple data units of the data acquisition module; Step 2: The data processing module converts the acquired 3D imaging data into a 3D anatomical model corresponding to the patient's torso and sorts each image data in time series; Step 3: The data processing module extracts regional features from the image data in the time series and generates a dynamic recovery curve; And the data processing module matches the dynamic recovery curve with the orthopedic database to assist doctors in judging whether the patient's recovery meets expectations; Step 4: During the operation, the data processing module connects the image data at the last time point in the time series with the three-dimensional anatomical model, maps the image data to the three-dimensional anatomical model, and obtains the preoperative model; At the same time, the navigation module integrates the preoperative model and obtains the patient's position information in real time during the operation, compares it with the preoperative model, and performs dynamic angle matching and correction.

Images