AI-based Rheumatoid Arthritis Risk Assessment Method and System

Abstract

The present invention relates to the technical field of rheumatoid arthritis assessment, and discloses an AI-based rheumatoid arthritis risk assessment method and system. A detection image of a target object is obtained to obtain a joint space narrowing characterization value of the detection image of the target object; during an evaluation period, the joint space narrowing characterization values of various detection images of the target object are respectively processed with a standard joint space narrowing characterization value to generate an evaluation status level signal; based on the evaluation status reminder signals of different levels of the target object during the evaluation period, the relevance of conditioning factors for the target object is analyzed to identify the influence degree of the conditioning factors on the target object during the evaluation period. The present invention completes the identification of the treatment effect of the target object according to the change situation of the joint space narrowing characterization value during the prediction period, and completes the visualization management of the target object during the evaluation period based on the evaluation status reminder signals of different levels of the target object during the evaluation period.

Description

Technical Field

[0001] The present invention relates to the technical field of rheumatoid arthritis assessment, and particularly to an AI-based rheumatoid arthritis risk assessment method and system in an embodiment. Background Art

[0002] Rheumatoid arthritis (RA) is a chronic, inflammatory autoimmune disease that mainly affects joints, causing joint swelling, pain, stiffness, and dysfunction. This disease usually affects multiple joints, especially the hands, feet, and knees, and the symptoms may recur. The pathogenesis of RA is complex and involves multiple factors such as genetics, environment, and immune system.

[0003] The treatment of rheumatoid arthritis requires a comprehensive approach, including drug treatment (such as non-steroidal anti-inflammatory drugs, immunosuppressants, and biological agents, etc.), physical therapy, and surgery when necessary.

[0004] In the prior art, the risk assessment methods for rheumatoid arthritis are all carried out during the diagnosis and treatment of patients, lacking the risk assessment of rheumatoid arthritis during the treatment process, which is not conducive to the targeted rehabilitation treatment of patients in the later stage. Especially during the identification of the degree of joint space stenosis in the detection image, it is based on the subjective judgment of doctors, which has certain limitations and lacks effective processing of the detection image. Summary of the Invention

[0005] The purpose of the present invention is to provide an AI-based rheumatoid arthritis risk assessment method and system. Based on the joint space stenosis characterization value corresponding to the detection image of the target object before the assessment period, the symptom conditioning degree of the target object during the assessment period is evaluated. That is, by obtaining the joint space stenosis characterization values of the detection images multiple times during the assessment period, and respectively comparing all the detected joint space stenosis characterization values during the assessment period with the standard joint space stenosis characterization value, the treatment effect of the target object is identified according to the change of the joint space stenosis characterization value during the prediction period, and based on the evaluation status reminder signals of different levels of the target object during the assessment period, the visualization management of the target object during the assessment period is completed.

[0006] The purpose of the present invention can be achieved through the following technical solutions:

[0007] An AI-based rheumatoid arthritis risk assessment method and system includes the following steps:

[0008] Obtain the detection image of the target object, process the detection image, and obtain the joint space stenosis characterization value of the detection image of the target object;

[0009] During the evaluation period, the joint space stenosis characterization values of multiple detection images of the target object are processed respectively with the standard joint space stenosis characterization value to generate an evaluation status level signal;

[0010] Among them, the evaluation status level signal includes an evaluation status improving signal and multi-level evaluation status reminder signals;

[0011] Based on the evaluation status improving signal, evaluate the treatment period of the target object;

[0012] Based on the different-level evaluation status reminder signals of the target object during the evaluation period, analyze the relevance of the conditioning factors of the target object, and identify the influence degree of the conditioning factors on the target object during the evaluation period.

[0013] As a further solution of the present invention: During the evaluation period, obtain the detection images of the target object;

[0014] And measure the joint space width of the detection images of the target object, and record the obtained joint space width value as the joint space stenosis characterization value of the target object.

[0015] As a further solution of the present invention: Obtain the detection images of the target object before the evaluation period, and record the joint space stenosis characterization value corresponding to the detection images before the evaluation period as the standard joint space stenosis characterization value, denoted as XZ0.

[0016] As a further solution of the present invention: During the evaluation period, record the joint space stenosis characterization value of the detection images of the target object as XZi, where i is the number of detection images;

[0017] Traverse the joint space stenosis characterization values of all the detection images of the target object during the evaluation period to obtain the maximum joint space stenosis characterization value XZmax and the minimum joint space stenosis characterization value XZmin;

[0018] If the maximum joint space stenosis characterization value XZmax during the evaluation period ≤ the standard joint space stenosis characterization value XZ0, generate an evaluation status level 1 reminder signal;

[0019] If the minimum joint space stenosis characterization value XZmin during the evaluation period > the standard joint space stenosis characterization value XZ0, generate an evaluation status improving signal.

[0020] As a further solution of the present invention: If the maximum joint space stenosis characterization value XZmax during the evaluation period > the standard joint space stenosis characterization value XZ0 ≥ the minimum joint space stenosis characterization value XZmin during the evaluation period;

[0021] Then process the joint space stenosis characterization values XZi of multiple detection images of the target object during the evaluation period;

[0022] Integrate the joint space stenosis characterization values of multiple detection images within the evaluation period to obtain a group of joint space stenosis characterization values;

[0023] Process the group of joint space stenosis characterization values according to the variance calculation formula to obtain the joint space stenosis characterization variance within the evaluation period, denoted as XZf;

[0024] Compare and process the joint space stenosis characterization variance XZf within the evaluation period with the preset joint space stenosis characterization variance threshold XZy.

[0025] As a further solution of the present invention: If the joint space stenosis characterization variance XZf within the evaluation period is ≥ the joint space stenosis characterization variance threshold XZy, generate a secondary evaluation status reminder signal;

[0026] When the joint space stenosis characterization variance XZf within the evaluation period < the joint space stenosis characterization variance threshold XZy, identify the joint space stenosis characterization value XZi within the evaluation period.

[0027] As a further solution of the present invention: If the joint space stenosis characterization value XZi within the evaluation period shows a linear continuous decrease, generate a tertiary evaluation status reminder signal;

[0028] If the joint space stenosis characterization value XZi within the evaluation period shows a linear continuous increase, generate a signal indicating an improving evaluation status;

[0029] If there is no linear relationship between the joint space stenosis characterization values XZi within the evaluation period, generate a quaternary evaluation status reminder signal.

[0030] As a further solution of the present invention: Divide the evaluation period into several evaluation time sub-units, and respectively obtain the joint physical rehabilitation dynamic values of the target object within each evaluation time sub-unit;

[0031] Establish a plane coordinate system, with the time of the evaluation period as the X-axis and the joint space stenosis characterization value as the Y-axis;

[0032] Plot the joint space stenosis characterization values corresponding to the target object's detection images in the plane coordinate system to obtain a joint space stenosis characterization curve;

[0033] Mark the area above the standard line of the joint space stenosis characterization curve as the positive joint space stenosis repair area;

[0034] Mark the area below the standard line of the joint space stenosis characterization curve as the negative joint space stenosis repair area;

[0035] Mark the evaluation time sub-units within the positive joint space stenosis repair area as positive time sub-units, and obtain the average joint physical rehabilitation dynamic value of the positive time sub-units;

[0036] Denote the evaluation time subunit located within the anti-repair area of joint space stenosis as the reverse time subunit, and obtain the dynamic mean value of joint physical rehabilitation of the reverse time subunit.

[0037] As a further solution of the present invention: If the difference between the dynamic mean value of joint physical rehabilitation of the forward time subunit and the dynamic mean value of joint physical rehabilitation of the reverse time subunit is within the preset requirement range, obtain a conditioning factor non-associated signal;

[0038] If the difference between the dynamic mean value of joint physical rehabilitation of the forward time subunit and the dynamic mean value of joint physical rehabilitation of the reverse time subunit is outside the preset requirement range, obtain a conditioning factor associated signal.

[0039] As a further solution of the present invention: An AI-based risk assessment system for rheumatoid arthritis, including:

[0040] An image analysis module, which is used to obtain the detection image of the target object, process the detection image, obtain the joint space stenosis characterization value of the target object detection image, and upload the joint space stenosis characterization value to the cloud control platform;

[0041] The evaluation and grading module receives the joint space stenosis characterization value transmitted by the cloud control platform. During the evaluation period, the evaluation and grading module processes the joint space stenosis characterization values of multiple detection images of the target object respectively with the standard joint space stenosis characterization value, generates an evaluation status level signal, and uploads the evaluation status level signal to the cloud control platform;

[0042] Among them, the evaluation status level signal includes an evaluation status improving signal and a multi-level evaluation status reminder signal;

[0043] The cycle prediction module receives the evaluation status improving signal transmitted by the cloud control platform. The cycle prediction module evaluates the treatment cycle of the target object based on the evaluation status improving signal;

[0044] The conditioning influence module receives the multi-level evaluation status reminder signal transmitted by the cloud control platform. The conditioning influence module analyzes the relevance of the conditioning factor of the target object based on the evaluation status reminder signals of different levels of the target object during the evaluation period, and identifies the influence degree of the conditioning factor on the target object during the evaluation period.

[0045] The beneficial effects of the present invention:

[0046] Based on the joint space narrowing characterization value corresponding to the target object's detection image before the evaluation period, the present invention evaluates the degree of symptom conditioning of the target object during the evaluation period (which can be understood as the treatment period of the target object). That is, by obtaining the joint space narrowing characterization values of multiple detection images during the evaluation period, and comparing all the detected joint space narrowing characterization values during the evaluation period with the standard joint space narrowing characterization value (the joint space narrowing characterization value corresponding to the target object's detection image before the evaluation period), the recognition of the treatment effect of the target object is completed according to the change of the joint space narrowing characterization value during the prediction period, and based on the evaluation status reminder signals of different levels of the target object during the evaluation period, the visualization management of the target object during the evaluation period is completed;

[0047] During the evaluation period of the present invention, when the joint space narrowing characterization values of all detection images of the target object are greater than and less than the standard joint space narrowing characterization value, that is, by establishing a joint space narrowing characterization curve, the area where the joint space narrowing characterization curve is above the standard line is recorded as the positive repair area of joint space narrowing, and the area where the joint space narrowing characterization curve is below the standard line is recorded as the negative repair area of joint space narrowing. By comparing and processing the conditioning factors (dynamic values of joint physical rehabilitation) in the positive repair area and negative repair area of joint space narrowing, the influence degree of the conditioning factors on the target object during the evaluation period is identified, so as to identify the conditioning influence of different conditioning factors on the target object, making the subsequent conditioning process of the target object more targeted. Brief Description of the Drawings

[0048] The following further describes the present invention with reference to the accompanying drawings.

[0049] Figure 1 is the flowchart of the AI-based rheumatoid arthritis risk assessment method according to an embodiment of the present invention;

[0050] Figure 2 is the flowchart of the conditioning factor association signal in the AI-based rheumatoid arthritis risk assessment method according to an embodiment of the present invention;

[0051] Figure 3 is the program block diagram of the AI-based rheumatoid arthritis risk assessment system according to an embodiment of the present invention. Detailed Embodiments

[0052] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention. Example 1

[0053] Please refer to Figure 1 As shown, the present invention is an AI-based risk assessment method for rheumatoid arthritis, including the following steps:

[0054] Obtain the detection image of the target object, process the detection image, and obtain the joint space narrowing characterization value of the target object's detection image;

[0055] During the evaluation period, the joint space narrowing characterization values of multiple detection images of the target object are respectively processed with the standard joint space narrowing characterization value to generate an evaluation status level signal;

[0056] Among them, the evaluation status level signal includes an evaluation status improving signal and multi-level evaluation status reminder signals;

[0057] Based on the evaluation status improving signal, evaluate the treatment cycle of the target object;

[0058] Based on the evaluation status reminder signals of different levels of the target object during the evaluation period, analyze the relevance of conditioning factors of the target object, and identify the influence degree of the conditioning factors on the target object during the evaluation period. Example 2

[0059] During the evaluation period, obtain the detection image of the target object;

[0060] And measure the joint space width of the detection image of the target object, and record the obtained joint space width value as the joint space narrowing characterization value of the target object;

[0061] Among them, the measurement process of the joint space width of the detection image is as follows:

[0062] Preprocess the detection image. Since the collected detection image may have problems such as noise and uneven illumination, the preprocessing of the detection image includes filtering, denoising, and enhancing contrast;

[0063] Perform gray processing on the detection image, obtain the pixel grid at the joint space width of the detection image, and obtain the joint space width of the detection image through statistical processing of the pixel grid.

[0064] Among them, there are no less than five detection images during the evaluation period;

[0065] Among them, the acquisition process of the detection image of the target object is as follows:

[0066] The probe emits ultrasonic waves to the detection position of the target object;

[0067] And receive the ultrasonic echo returned from the detection object through the probe to obtain the ultrasonic echo signal / data;

[0068] Perform beamforming processing on the ultrasonic echo signal / data to obtain the ultrasonic echo signal / data after beamforming. After normalization, logarithmic compression, and dynamic range limiting processing, a detection image of the target object is obtained.

[0069] During the evaluation period, record the joint space stenosis characterization value of the detection image of the target object as XZi, where i is the number of detection images, and i = 1, ……, n;

[0070] Obtain the detection image of the target object before the evaluation period, and record the joint space stenosis characterization value corresponding to the detection image before the evaluation period as the standard joint space stenosis characterization value, denoted as XZ0;

[0071] Traverse the joint space stenosis characterization values of all detection images of the target object during the evaluation period to obtain the maximum joint space stenosis characterization value XZmax and the minimum joint space stenosis characterization value XZmin;

[0072] If the maximum joint space stenosis characterization value XZmax ≤ the standard joint space stenosis characterization value XZ0 during the evaluation period, it indicates that the symptoms (rheumatoid arthritis) at the detection position of the target object gradually deteriorate during the evaluation period, and the joint space stenosis characterization value gradually decreases, generating a first-level reminder signal for the evaluation status;

[0073] If the minimum joint space stenosis characterization value XZmin > the standard joint space stenosis characterization value XZ0 during the evaluation period, it indicates that the symptoms (rheumatoid arthritis) at the detection position of the target object gradually improve during the evaluation period, and the joint space stenosis characterization value gradually increases, generating a signal indicating improvement in the evaluation status;

[0074] If the maximum joint space stenosis characterization value XZmax > the standard joint space stenosis characterization value XZ0 ≥ the minimum joint space stenosis characterization value XZmin during the evaluation period;

[0075] Then process the joint space stenosis characterization values XZi of multiple detection images of the target object during the evaluation period;

[0076] Integrate the joint space stenosis characterization values of multiple detection images during the evaluation period to obtain a joint space stenosis characterization value group;

[0077] Process the joint space stenosis characterization value group according to the variance calculation formula to obtain the joint space stenosis characterization variance during the evaluation period, denoted as XZf;

[0078] Denote the joint space stenosis characterization variance threshold during the evaluation period as XZy;

[0079] Compare and process the variance XZf of the joint space stenosis characterization during the evaluation period with the preset variance threshold XZy of the joint space stenosis characterization;

[0080] When the variance XZf of the joint space stenosis characterization during the evaluation period is greater than or equal to the variance threshold XZy of the joint space stenosis characterization, it indicates that the fluctuation of the joint space stenosis characterization value during the evaluation period is large, that is, the conditioning effect of the symptoms (rheumatoid arthritis) at the detection position during the evaluation period is poor, and a secondary reminder signal for the evaluation status is generated;

[0081] When the variance XZf of the joint space stenosis characterization during the evaluation period is less than the variance threshold XZy of the joint space stenosis characterization, it indicates that the fluctuation of the joint space stenosis characterization value during the evaluation period is small, and identify the joint space stenosis characterization value XZi during the evaluation period;

[0082] If the joint space stenosis characterization value XZi during the evaluation period shows a linear continuous decrease, it indicates that the degree of joint space stenosis gradually worsens during the evaluation period, that is, during the evaluation period, the conditioning effect of the symptoms (rheumatoid arthritis) at the detection position is not ideal and gradually worsens, and a tertiary reminder signal for the evaluation status is generated;

[0083] If the joint space stenosis characterization value XZi during the evaluation period shows a linear continuous increase, it indicates that the degree of joint space stenosis gradually decreases during the evaluation period, that is, during the evaluation period, the conditioning effect of the symptoms (rheumatoid arthritis) at the detection position is ideal and gradually improves, and a signal indicating that the evaluation status is improving is generated;

[0084] If there is no linear relationship between the joint space stenosis characterization values XZi during the evaluation period, it indicates that the degree of joint space stenosis fluctuates little during the evaluation period, that is, during the evaluation period, the conditioning effect of the symptoms (rheumatoid arthritis) at the detection position is not ideal and the symptoms do not change significantly, and a quaternary reminder signal for the evaluation status is generated.

[0085] It should be noted that: the severity of the symptoms at the detection position of the target object corresponding to the primary reminder signal for the evaluation status is higher than the severity of the symptoms at the detection position corresponding to the secondary reminder signal for the evaluation status;

[0086] The severity of the symptoms at the detection position of the target object corresponding to the secondary reminder signal for the evaluation status is higher than the severity of the symptoms at the detection position corresponding to the tertiary reminder signal for the evaluation status;

[0087] The severity of the symptoms at the detection position of the target object corresponding to the tertiary reminder signal for the evaluation status is higher than the severity of the symptoms at the detection position corresponding to the quaternary reminder signal for the evaluation status.

[0088] The technical solution of the present invention is as follows: Based on the joint space stenosis characterization value corresponding to the target object's detection image before the evaluation period, the degree of symptom conditioning of the target object during the evaluation period (which can be understood as the treatment period of the target object) is evaluated. That is, by obtaining the joint space stenosis characterization values of multiple detection images during the evaluation period, and comparing and processing all the detected joint space stenosis characterization values during the evaluation period with the standard joint space stenosis characterization value (the joint space stenosis characterization value corresponding to the target object's detection image before the evaluation period), the recognition of the treatment effect of the target object is completed according to the change of the joint space stenosis characterization value within the prediction period;

[0089] And based on the evaluation status reminder signals of different levels of the target object during the evaluation period, the visual management of the target object during the evaluation period is completed. Embodiment 3

[0090] Based on the evaluation status improvement signal of the target object during the evaluation period, the treatment period of the target object is evaluated;

[0091] Specifically:

[0092] During the evaluation period, the difference between the joint space stenosis characterization values of all the detection images of the target object and the standard joint space stenosis characterization value is calculated respectively, and the sum of the obtained differences is averaged to obtain the joint space stenosis repair value of the evaluation period;

[0093] That is, through the formula The joint space stenosis repair value XF of the evaluation period is calculated;

[0094] Obtain the normal joint space stenosis characterization value of the target object;

[0095] The difference between the normal joint space stenosis characterization value and the standard joint space stenosis characterization value is calculated to obtain the joint space stenosis repair interval value;

[0096] The ratio of the joint space stenosis repair interval value to the joint space stenosis repair value is calculated to obtain the treatment period of the target object;

[0097] Among them, the normal joint space stenosis characterization value is the normal joint space stenosis characterization value corresponding to the normal joint when the target object has no symptoms, which is an empirical value.

[0098] The technical solution of the present invention is as follows: when the evaluation status of the target object shows an improvement during the evaluation period, the repair degree of the joints of the target object during the evaluation period is obtained. Specifically, the difference between the joint space stenosis characterization values of all the detection images of the target object and the standard joint space stenosis characterization value is calculated respectively, and the sum of the obtained differences is averaged to obtain the joint space stenosis repair value during the evaluation period. Then, the joint space stenosis repair analysis value (total repair value) of the target object is processed with the joint space stenosis repair value during the evaluation period to obtain the estimated number of repair cycles. Example 4

[0099] Based on the first-level reminder signal of the evaluation status of the target object during the evaluation period, it indicates that the conditioning of the target object during the evaluation period has no effect, and the symptoms of the target object are becoming more and more serious, and the treatment plan for the target object needs to be adjusted. Example 5

[0100] Based on the reminder signals of different levels (second level, third level or fourth level) of the evaluation status of the target object during the evaluation period, the relevance of the conditioning factors of the target object is analyzed;

[0101] Among them, the conditioning factors of the target object include the amount of drug taken (methotrexate), joint physical rehabilitation data (including but not limited to hot compress and acupuncture), etc.;

[0102] In this embodiment, an example of the joint physical rehabilitation dynamic value of the target object is given;

[0103] The evaluation period is divided into several evaluation time sub-units, and the joint physical rehabilitation dynamic values of the target object in each evaluation time sub-unit are obtained respectively;

[0104] Among them, the process of obtaining the joint physical rehabilitation dynamic value is as follows:

[0105] Obtain the number of physical rehabilitation (joint activity training) times and the duration of each rehabilitation within the evaluation time sub-unit;

[0106] Obtain the physical rehabilitation frequency within the evaluation time sub-unit;

[0107] Sum up the duration of each physical rehabilitation to obtain the total physical rehabilitation duration, and calculate the ratio of the total physical rehabilitation duration to the duration of the evaluation time sub-unit to obtain the physical rehabilitation time ratio;

[0108] Multiply the physical rehabilitation frequency by the physical rehabilitation time ratio to obtain the joint physical rehabilitation dynamic value.

[0109] Establish a plane coordinate system, with the time of the evaluation period as the X-axis and the joint space stenosis characterization value as the Y-axis;

[0110] In chronological order, in a plane coordinate system, plot the joint space stenosis characterization values corresponding to the target object detection images. In the coordinate system, connect all the joint space stenosis characterization values with a smooth curve from left to right to obtain a joint space stenosis characterization curve;

[0111] In the plane coordinate system, draw a standard line parallel to the X-axis with the standard joint space stenosis characterization value;

[0112] Mark the area above the standard line of the joint space stenosis characterization curve as the positive joint space stenosis repair area;

[0113] Mark the area below the standard line of the joint space stenosis characterization curve as the negative joint space stenosis repair area;

[0114] Mark the evaluation time sub-units within the positive joint space stenosis repair area as positive time sub-units;

[0115] Obtain the joint physical rehabilitation dynamic values of all positive time sub-units, sum up and average the joint physical rehabilitation dynamic values of all positive time sub-units to obtain the average joint physical rehabilitation dynamic value of the positive time sub-units;

[0116] Mark the evaluation time sub-units within the negative joint space stenosis repair area as negative time sub-units;

[0117] Obtain the joint physical rehabilitation dynamic values of all negative time sub-units, sum up and average the joint physical rehabilitation dynamic values of all negative time sub-units to obtain the average joint physical rehabilitation dynamic value of the negative time sub-units;

[0118] Refer to Figure 2 , if the difference between the average joint physical rehabilitation dynamic value of the positive time sub-units and the average joint physical rehabilitation dynamic value of the negative time sub-units is within the preset requirement range, it indicates that the poor conditioning effect on the target object within the negative time sub-units is not caused by the joint physical rehabilitation data, and a conditioning factor non-associated signal is obtained;

[0119] If the difference between the average joint physical rehabilitation dynamic value of the positive time sub-units and the average joint physical rehabilitation dynamic value of the negative time sub-units is outside the preset requirement range, it indicates that the poor conditioning effect on the target object within the negative time sub-units is caused by the joint physical rehabilitation data, and a conditioning factor associated signal is obtained.

[0120] Based on the conditioning factor non-associated signal, conduct a correlation analysis on other conditioning factors of the target object during the evaluation period. Other conditioning factors include, but are not limited to, the dosage of anti-rheumatic drugs, the usage time of anti-rheumatic drugs, etc., and complete the evaluation and analysis of the target object in multiple dimensions.

[0121] Based on the conditioning factor correlation signal, with the joint physical rehabilitation dynamic mean of the forward time subunit as a reference, the target object undergoes joint physical rehabilitation during the evaluation period.

[0122] The technical solution of the present invention is as follows: During the evaluation period, when the joint space narrowing characterization values of all the detection images of the target object are greater than the standard joint space narrowing characterization value and less than the standard joint space narrowing characterization value, that is, by establishing a joint space narrowing characterization curve, the area where the joint space narrowing characterization curve is above the standard line is recorded as the positive repair area of joint space narrowing, and the area where the joint space narrowing characterization curve is below the standard line is recorded as the negative repair area of joint space narrowing. By comparing and processing the conditioning factors (joint physical rehabilitation dynamic values) in the positive repair area of joint space narrowing and the negative repair area of joint space narrowing, the influence degree of the conditioning factors on the target object during the evaluation period is identified, so as to identify the conditioning effects of different conditioning factors on the target object, making the subsequent conditioning process of the target object more targeted. Example 6

[0123] Please refer to Figure 3 As shown, the present invention is an AI-based risk assessment system for rheumatoid arthritis, including an image analysis module, an evaluation and grading module, a cycle prediction module, a conditioning influence module, and a cloud control platform;

[0124] The image analysis module is used to obtain the detection images of the target object, process the detection images, and obtain the joint space narrowing characterization values of the detection images of the target object. The joint space narrowing characterization values are uploaded to the cloud control platform;

[0125] The evaluation and grading module receives the joint space narrowing characterization values transmitted by the cloud control platform. During the evaluation period, the evaluation and grading module processes the joint space narrowing characterization values of multiple detection images of the target object respectively with the standard joint space narrowing characterization value, generates an evaluation status level signal, and uploads the evaluation status level signal to the cloud control platform;

[0126] Among them, the evaluation status level signal includes an evaluation status improvement signal and a multi-level evaluation status reminder signal;

[0127] The cycle prediction module receives the evaluation status improvement signal transmitted by the cloud control platform. Based on the evaluation status improvement signal, the cycle prediction module evaluates the treatment cycle of the target object;

[0128] The conditioning influence module receives the multi-level evaluation status reminder signal transmitted by the cloud control platform. Based on the different-level evaluation status reminder signals of the target object during the evaluation period, the conditioning influence module analyzes the conditioning factor correlation of the target object and identifies the influence degree of the conditioning factors on the target object during the evaluation period.

[0129] The above has described in detail an embodiment of the present invention, but the above content is only a preferred embodiment of the present invention and cannot be considered as limiting the scope of implementation of the present invention. All equivalent changes and improvements made according to the scope of the application of the present invention should still fall within the scope covered by the patent of the present invention.

Claims

1. An AI-based rheumatoid arthritis risk assessment method, characterized in that: The following steps are involved: During the evaluation period, a detection image of the target object is obtained; and measuring the joint gap width of the detection image of the target object, and recording the obtained joint gap width value as the joint gap narrowing characterization value of the target object; Acquire a detection image of the target object before the evaluation period, and record the joint space stenosis characterization value corresponding to the detection image before the evaluation period as a standard joint space stenosis characterization value, recorded as XZ0; During the evaluation period, the joint space stenosis characterization values ​​of the target object's various detection images are compared with the standard joint space stenosis characterization values ​​to generate an evaluation state level signal; Among them, the evaluation status level signal includes an evaluation status improvement signal and a multi-level evaluation status reminder signal; Based on the positive signal of the evaluation status, the treatment cycle of the target object is evaluated; Based on the target object's assessment status reminder signals at different levels during the assessment cycle, the correlation of the target object's conditioning factors is analyzed to identify the degree of influence of the conditioning factors on the target object during the assessment cycle; The analysis of the correlation of conditioning factors of the target object includes: The evaluation period is divided into a number of evaluation time sub-units, and the joint physical rehabilitation dynamic value of the target object in each evaluation time sub-unit is obtained respectively; Among them, the process of obtaining the dynamic value of joint physical rehabilitation is as follows: Obtain the number of physical rehabilitation times and the duration of each rehabilitation session within the assessment time subunit; The frequency of physical rehabilitation within the assessment time subunit was obtained; The duration of each physical rehabilitation session is summed up to obtain the total duration of physical rehabilitation, and the total duration of physical rehabilitation is calculated to be proportional to the duration of the evaluation time subunit to obtain the physical rehabilitation time ratio; The physical rehabilitation frequency is multiplied by the physical rehabilitation time ratio to obtain the joint physical rehabilitation dynamic value; Establish a plane coordinate system, with the time of the evaluation cycle as the X-axis and the value representing the narrowing of the joint space as the Y-axis; In chronological order, the joint space stenosis characterization values ​​corresponding to the target object detection image are plotted in the plane coordinate system, and all the joint space stenosis characterization values ​​are connected from left to right in the coordinate system with a smooth curve to obtain a joint space stenosis characterization curve; In the plane coordinate system, a standard line parallel to the X-axis is drawn with the standard joint space narrowing characterization value; The area where the curve characterizing joint space stenosis is above the standard line is recorded as the joint space stenosis positive repair area; The area where the curve representing joint space stenosis is below the standard line is recorded as the anti-repair area of ​​joint space stenosis; The evaluation time subunit located in the positive repair area of ​​the joint space narrowing was recorded as the positive time subunit; Acquire the joint physical rehabilitation dynamic values ​​of all forward time subunits, sum and average the joint physical rehabilitation dynamic values ​​of all forward time subunits, and obtain the joint physical rehabilitation dynamic average of the forward time subunits; The evaluation time subunit located in the anti-repair area of ​​joint space narrowing was recorded as the reverse time subunit; Acquire the joint physical rehabilitation dynamic values ​​of all reverse time subunits, sum and average the joint physical rehabilitation dynamic values ​​of all reverse time subunits, and obtain the joint physical rehabilitation dynamic average of the reverse time subunits; If the difference between the joint physical rehabilitation dynamic mean value of the forward time subunit and the joint physical rehabilitation dynamic mean value of the reverse time subunit is within the preset requirement range, it means that the poor conditioning effect on the target object in the reverse time subunit is not caused by the joint physical rehabilitation data, and a conditioning factor non-correlation signal is obtained; If the difference between the dynamic mean of joint physical rehabilitation of the forward time subunit and the dynamic mean of joint physical rehabilitation of the reverse time subunit is outside the preset requirement range, it means that the poor conditioning effect on the target object in the reverse time subunit is caused by the joint physical rehabilitation data, and a conditioning factor association signal is obtained.

2. The AI-based rheumatoid arthritis risk assessment method according to claim 1, characterized in that: The joint space stenosis characterization value of the target object detection image within the evaluation period is recorded as XZi, where i is the number of detection images; Traversing the joint space stenosis representation values ​​of all detection images of the target object within the evaluation period, and obtaining the maximum joint space stenosis representation value XZmax and the minimum joint space stenosis representation value XZmin; If the maximum joint space stenosis characterization value XZmax within the evaluation period is ≤ the standard joint space stenosis characterization value XZ0, a first-level evaluation status reminder signal is generated; If the minimum joint space narrowing characterization value XZmin within the evaluation period is greater than the standard joint space narrowing characterization value XZ0, a signal that the evaluation status is improving is generated.

3. The AI-based rheumatoid arthritis risk assessment method according to claim 2, characterized in that: If the maximum joint space narrowing characterization value XZmax during the evaluation period is greater than the standard joint space narrowing characterization value XZ0 ≥ the minimum joint space narrowing characterization value XZmin during the evaluation period; Integrating the joint space stenosis characterization values ​​of multiple detection images within an evaluation period to obtain a joint space stenosis characterization value group; The joint space stenosis characterization value group was processed according to the variance calculation formula to obtain the joint space stenosis characterization variance within the evaluation period, which was recorded as XZf; The variance XZf representing the joint space narrowing during the evaluation period is compared with the preset variance threshold XZy representing the joint space narrowing.

4. The AI-based rheumatoid arthritis risk assessment method according to claim 3, characterized in that: If the variance XZf representing the narrowing of the joint space is greater than or equal to the variance threshold XZy representing the narrowing of the joint space during the evaluation period, a secondary reminder signal of the evaluation status is generated.

5. The AI-based rheumatoid arthritis risk assessment method according to claim 4, characterized in that: When the variance XZf representing the narrowing of the joint space in the evaluation period is less than the variance threshold XZy representing the narrowing of the joint space, the value XZi representing the narrowing of the joint space in the evaluation period is identified; If the joint space stenosis characterization value XZi decreases linearly and continuously during the evaluation period, a third-level reminder signal of the evaluation status is generated; If the joint space narrowing characterization value XZi increases linearly and continuously during the evaluation period, a signal that the evaluation status is improving is generated; If the joint space stenosis characterization value XZi within the evaluation period has no linear relationship, a fourth-level reminder signal of the evaluation status is generated.

6. An AI-based rheumatoid arthritis risk assessment system, the system being used to execute the AI-based rheumatoid arthritis risk assessment method according to any one of claims 1 to 5, characterized in that: include: An image analysis module is used to obtain a detection image of the target object, process the detection image, obtain a joint space stenosis representation value of the target object detection image, and upload the joint space stenosis representation value to the cloud management and control platform; The evaluation and grading module receives the joint space stenosis characterization value transmitted by the cloud management and control platform. During the evaluation period, the evaluation and grading module processes the joint space stenosis characterization values ​​of the target object's various detection images with the standard joint space stenosis characterization value, generates an evaluation status level signal, and uploads the evaluation status level signal to the cloud management and control platform; Among them, the evaluation status level signal includes an evaluation status improvement signal and a multi-level evaluation status reminder signal; The cycle estimation module receives the evaluation status improvement signal transmitted by the cloud management and control platform, and the cycle estimation module evaluates the treatment cycle of the target object based on the evaluation status improvement signal; The conditioning impact module receives multi-level assessment status reminder signals transmitted by the cloud management and control platform. Based on the assessment status reminder signals of different levels of the target object during the assessment cycle, the conditioning impact module analyzes the correlation of the conditioning factors of the target object and identifies the degree of influence of the conditioning factors on the target object during the assessment cycle.

Images