An AI-based intelligent management system for multi-disease chronic disease information

Abstract

The present invention discloses an AI-based intelligent management system for multi-disease chronic disease information, which relates to the field of medical information technology. The system, equipped with a laser response capture module and a multidimensional data processing module, can collect RGB image datasets and infrared thermal map datasets of the skin area at set time points after laser treatment. Based on an AI image processing model, the images are standardized and the white spot area is segmented. Morphological parameters such as the area of ​​the white spot, edge curvature kurtosis, and tangential direction disturbance, as well as thermal parameters such as temperature gradient recovery value, temperature change curvature, and regional heat flux density, are then extracted. The system calculates the structural convergence tension index (STCI) and the thermal response consistency mapping index (HRCI) through the coordinated operation of the structural convergence analysis module and the thermal recovery analysis module. This system comprehensively measures the skin response process from both structural and thermal dimensions, effectively compensating for the inaccurate assessment caused by traditional reliance on visual observation and subjective judgment.

Description

Technical Field

[0001] The present invention relates to the field of medical information technology, and specifically to an AI-based intelligent management system for multi-disease chronic disease information. Background Art

[0002] With the continuous integration of artificial intelligence (AI) technology, image processing algorithms and medical big data platforms, smart medical systems have become one of the important development directions of the modern medical industry. In this direction, intelligent management technology for chronic diseases is particularly critical because it has natural needs in long-term treatment, personalized intervention and remote follow-up. In this field, chronic skin diseases are characterized by obvious manifestations and slow changes, especially in pigmentary dyschromias such as vitiligo. 308nm excimer laser treatment is a mainstream physical treatment method widely used in clinical practice. Because of its stable wavelength, focused energy and clear target, it can effectively induce melanocyte migration, proliferation and pigment synthesis, making it the preferred treatment option for patients with moderate to severe vitiligo. However, this treatment usually requires multiple cycles and multiple irradiations to show results, and the judgment of efficacy relies heavily on the doctor's subjective assessment of the changes in the appearance of white spots after treatment.

[0003] At present, during laser treatment, doctors mainly rely on visual inspection of external phenomena such as the reduction of white spots, pigment reflow and erythema disappearance after each treatment to judge the efficacy. However, this method is obviously subjective and has a lag, and cannot accurately distinguish between temporary reactions and actual recovery. In addition, although some hospitals are equipped with image acquisition equipment such as dermatoscopes, RGB cameras or thermal imagers, these devices are only used as "static auxiliary tools" and are not deeply integrated with the back-end analysis system. There is a lack of dynamic linkage analysis models of images, temperature and structure. What's more, the key time series evaluation logic of the "treatment response curve" is missing, so it is impossible to build a systematic efficacy prediction mechanism and intelligent adjustment recommendation mechanism;

[0004] The root cause of the above situation lies in the lack of unified data collection standards in the existing treatment system, the lack of AI model processing paths for image and temperature data, and the lack of a closed-loop logic between treatment results and structural morphology and thermal recovery. This defect makes it impossible for doctors to accurately judge whether patients who do not respond to treatment are experiencing "delayed recovery" or "ineffective treatment", and it is even more impossible to formulate dynamic treatment strategies based on the degree of recovery. This will lead to: first, the treatment course is passively extended and the burden on patients is increased; second, the treatment effect is misjudged, and treatment plans may be frequently switched; third, some patients clearly have early signs of recovery but are not caught, missing the optimal time point for adjustment. Ultimately, this affects the individualization, scientific nature, and predictability of treatment. Summary of the Invention

[0005] In response to the shortcomings of the existing technology, the present invention provides an AI-based intelligent management system for multi-disease chronic disease information, which solves the problems mentioned in the background technology.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: including a laser response capture module, a multi-dimensional data processing module, a structural convergence analysis module, a thermal recovery analysis module and a comprehensive analysis module;

[0007] The laser response capture module collects the RGB image dataset and infrared thermal map dataset of the skin area after laser treatment in real time by using a visual sensor, and transmits the RGB image dataset and infrared thermal map dataset to the information management system;

[0008] The multidimensional data processing module preprocesses the RGB image dataset and the infrared thermal image dataset by using an AI image processing model to obtain an RGB white spot area image and an infrared white spot area image, and extracts features from the RGB white spot area image to obtain standardized white spot data;

[0009] The structural convergence analysis module calculates and outputs the structural convergence tension index (STCI) based on the standardized vitiligo data, sets a convergence interval threshold for preliminary comparative evaluation, and triggers skin thermal recovery analysis based on the preliminary comparative evaluation results;

[0010] The thermal recovery analysis module extracts thermal recovery features from the infrared white spot area image to obtain standardized thermal recovery data, and calculates and outputs the thermal response consistency mapping index HRCI based on the standardized thermal recovery data;

[0011] The comprehensive analysis module summarizes and calculates the thermal response consistency mapping index HRCI and the structural convergence tension index STCI, outputs the comprehensive matching index SHMI, and sets the recovery interval threshold and the comprehensive matching index SHMI for secondary comparative evaluation.

[0012] Preferably, the laser response capturing module includes a white spot image capturing unit and an image data transmitting unit;

[0013] The vitiligo image capture unit uses a visual sensor to collect RGB image data sets and infrared thermal image data sets of the skin area after laser treatment in real time by setting image collection time points after laser treatment.

[0014] The image acquisition time points were set at 12 hours, 48 ​​hours, and 96 hours after each laser treatment;

[0015] The visual sensor includes an RGB visual sensor and a thermal imaging visual sensor;

[0016] The image data transmission unit uses the built-in wireless communication module of the visual sensor and sets up a 5G communication network to wirelessly connect the visual sensor with the information management system, and transmits the RGB image data set and the infrared thermal map data set to the information management system in real time.

[0017] Preferably, the multi-dimensional data processing module includes an AI embedding unit and an RGB image processing unit;

[0018] The AI ​​embedding unit embeds an AI image processing model in the information management system and inputs the real-time received RGB image dataset and infrared thermal image dataset into the AI ​​image processing model for preprocessing;

[0019] The AI ​​image processing model includes an image standardization preprocessing module, an RGB feature extraction module, and a thermal recovery feature extraction module;

[0020] The image standardization preprocessing module loads OpenCV image processing technology into the AI ​​image processing model to standardize the RGB image dataset and the infrared thermal map dataset, and performs size unification, image alignment, intensity normalization and color correction on the RGB image dataset and the infrared thermal map dataset;

[0021] The white spot areas in all images in the RGB image dataset and infrared thermal map dataset after image standardization preprocessing are outlined using medical image annotation software, and an AI segmentation model is constructed using the semantic segmentation network U-Net. A large number of samples with the white spot area outlines are input into the AI ​​segmentation model to train the AI ​​segmentation model, and the real-time RGB image dataset and infrared thermal map dataset are input into the AI ​​segmentation model to segment the RGB image dataset and the infrared thermal map dataset, and extract the RGB white spot area images and the infrared white spot area images.

[0022] Preferably, the RGB image processing unit transmits the RGB white spot area image obtained by the AI ​​image processing model to the RGB feature extraction module to perform feature extraction to obtain white spot morphological data;

[0023] The white spot morphological data includes the white spot area Sb(t) at time t, the edge curvature distribution kurtosis Pc(t) at time t, and the tangential direction disturbance amplitude Ttan(t) at time t;

[0024] The white spot morphology data were normalized using Z-score to eliminate the dimensions of all parameters in the white spot morphology data and unify the units.

[0025] Preferably, the structural convergence analysis module includes a white spot structural convergence analysis unit and a convergence evaluation unit;

[0026] The white spot structure convergence analysis unit extracts white spot morphological data and calculates and outputs a structural convergence tension index STCI to measure the convergence of the white spot boundary;

[0027] The structural convergence tension index STCI is calculated and output by the following algorithm formula:

[0028] ;

[0029] Where n represents the total number of image acquisition time points, cos represents the cosine function, represents the standard deviation of the curvature distribution, d represents the calculus function, dt represents the time calculus function, Represents the direction similarity factor.

[0030] Preferably, the convergence evaluation unit sets a convergence interval threshold, which includes a first convergence interval threshold F1 and a second convergence interval threshold F2, and performs a preliminary comparative evaluation on the structural convergence tension index STCI obtained in real time and the convergence interval threshold, to judge the recovery of the structure level after the vitiligo laser treatment, and triggers the skin thermal recovery analysis based on the preliminary comparative evaluation results. The specific evaluation contents are as follows;

[0031] When the structural convergence tension index STCI is greater than the first convergence interval threshold F1, it indicates that the white spot structure is stably converging, that is, the white spot area continues to shrink, the boundary is smooth and stable, and the shrinkage direction is consistent, and no adjustment is performed at this time;

[0032] When the second convergence interval threshold F2 < structural convergence tension index STCI ≤ the first convergence interval threshold F1, it indicates that there is a fluctuation response in the white spot structure contraction, and it is prompted to execute the image acquisition time point again for acquisition and analysis;

[0033] When the structural convergence tension index STCI ≥ the second convergence interval threshold F2, it indicates that the white spot structure is unresponsive, and the skin thermal recovery analysis is triggered.

[0034] Preferably, the thermal recovery analysis module includes a thermal recovery feature extraction unit and a thermal response consistency analysis unit;

[0035] The thermal recovery feature extraction unit extracts the infrared white spot area image processed by the image standardization preprocessing module after triggering the skin thermal recovery analysis, inputs the image into the thermal recovery feature extraction module in the AI ​​image processing model, extracts the thermal recovery features of the infrared white spot area image, and performs Z-score normalization on the feature extraction results to eliminate the dimensional influence of all extraction results, thereby obtaining standardized thermal recovery data;

[0036] The standardized thermal recovery data includes the temperature gradient recovery value ΔTr(t) at time t, the local curvature Kf(t) of the temperature recovery curve at time t, and the regional heat flux reconstruction density Rs(t) at time t.

[0037] Preferably, the thermal response consistency analysis unit extracts standardized thermal recovery data, calculates and outputs a thermal response consistency mapping index HRCI, and analyzes the coupling strength between thermal response and structural trend;

[0038] The thermal response consistency mapping index HRCI is calculated and output by the following algorithm formula:

[0039] ;

[0040] Where HRCI(t) represents the thermal response consistency mapping index at time t, T represents the image acquisition time interval, log represents the logarithmic function, d represents the calculus function, and dt represents the time calculus function.

[0041] Preferably, the comprehensive analysis module includes a structural and thermal resonance analysis unit and a comprehensive evaluation unit;

[0042] The structural and thermal resonance analysis unit calculates and outputs a comprehensive matching index SHMI based on the acquired thermal response consistency mapping index HRCI and structural convergence tension index STCI, thereby measuring the morphological convergence trend and thermal metabolic coupling behavior of the current vitiligo.

[0043] The comprehensive matching index SHMI is calculated and output by the following algorithm formula:

[0044] ;

[0045] Where SHMI(t) represents the comprehensive matching index at time t, Represents the average value of the local curvature of the temperature recovery curve.

[0046] Preferably, the comprehensive evaluation unit performs a secondary comparative evaluation on the current white spot area by setting a recovery interval threshold and the comprehensive matching index SHMI obtained in real time, judges the recovery status of the current white spot area, and divides the recovery level based on the secondary comparative evaluation result, and performs corresponding information management based on the divided recovery level;

[0047] The recovery interval threshold includes a first recovery interval threshold Q1 and a second recovery interval threshold Q2;

[0048] The specific assessment contents are as follows;

[0049] When the comprehensive matching index SHMI is greater than the first recovery interval threshold Q1, it is classified as recovery level A. At this time, the current governance plan is maintained without intervention;

[0050] When the second recovery interval threshold Q2 is less than the comprehensive matching index SHMI ≤ the first recovery interval threshold Q1, it is classified as recovery level B. At this time, it is prompted to increase the frequency of laser treatment by 50%, and a melanin nutrition reminder is sent to remind you to supplement tyrosinase;

[0051] When the comprehensive matching index SHMI ≤ the second recovery interval threshold Q2, it is classified as recovery level C. At this time, a re-examination prompt for the current white spot area is generated, and the laser parameters are doubled.

[0052] This invention provides an AI-based intelligent management system for multi-disease chronic disease information. It has the following beneficial effects:

[0053] (1) The system is equipped with a laser response capture module and a multi-dimensional data processing module. It can collect RGB image data sets and infrared thermal map data sets of the skin area at set time points after laser treatment, and perform standardization processing on the images and segment the white spot area based on the AI ​​image processing model, and then extract morphological parameters such as the area of ​​the white spot, edge curvature kurtosis and tangential direction disturbance, as well as thermal parameters such as temperature gradient recovery value, temperature change curvature and regional heat flux density. The system works in conjunction with the structural convergence analysis module and the thermal recovery analysis module to calculate the structural convergence tension index STCI and the thermal response consistency mapping index HRCI, comprehensively measuring the skin response process from the two dimensions of structure and heat, effectively compensating for the inaccurate evaluation caused by traditional reliance on naked eye observation and subjective judgment, and significantly improving the objectivity and continuity of efficacy judgment.

[0054] (2) The system uses a comprehensive analysis module to fuse the structural convergence tension index STCI and the thermal response consistency mapping index HRCI to form the structural-thermal resonance matching index SHMI, and sets the recovery interval threshold to achieve a secondary assessment of the recovery of the white spot area. The system can be divided into three levels according to the SHMI level: A recovery level, significant recovery, B recovery level, moderate response and C recovery level, low response, and implement differentiated management strategies based on the level results: maintain the original treatment plan, adjust the laser frequency or prompt for review and mode switching. This mechanism realizes the feedback closed-loop control mechanism of the treatment response, so that the AI ​​model not only has the ability to passively judge, but also can actively infer and execute treatment recommendations, thereby improving the targeting and dynamic regulation capabilities of vitiligo treatment.

[0055] (3) The system covers three time points through the image acquisition and analysis process. Through time series data standardization, AI feature extraction and indicator calculation modules, the system can perform differential analysis, curve fitting and statistical modeling on the structural and thermal evolution characteristics of the white spot area within the same treatment cycle. At the same time, the system uses AI deep processing methods such as Z-score standardization and graph convolutional network (GCN) to ensure the comparability of parameters between different acquisition time points, and the output parameters have time consistency and dynamic tracking capabilities. The system is not only suitable for single efficacy evaluation, but can also perform efficacy trend comparison and response pattern cluster analysis based on historical data, thereby providing a high-quality data foundation for long-term efficacy monitoring of chronic diseases, large-sample AI model training and treatment strategy optimization. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 This is a flow chart of an AI-based intelligent management system for multi-disease chronic disease information in the present invention;

[0057] Figure 2 It is a trend chart of the evolution of standardized thermal recovery data;

[0058] Figure 3 This is a trend chart of the evolution of white spot morphology data. DETAILED DESCRIPTION

[0059] The following will provide a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0060] Example 1

[0061] See also Figure 1 、 Figure 2 and Figure 3 The present invention provides an AI-based intelligent management system for multi-disease chronic disease information. To achieve the above purpose, the present invention is implemented through the following technical solutions: including a laser response capture module, a multi-dimensional data processing module, a structural convergence analysis module, a thermal recovery analysis module and a comprehensive analysis module;

[0062] The laser response capture module uses a visual sensor to collect the RGB image dataset and infrared thermal map dataset of the skin area after laser treatment in real time, and transmits the RGB image dataset and infrared thermal map dataset to the information management system;

[0063] The multi-dimensional data processing module pre-processes the RGB image dataset and the infrared thermal image dataset by using an AI image processing model to obtain RGB white spot area images and infrared white spot area images, and extracts features from the RGB white spot area images to obtain standardized white spot data;

[0064] The structural convergence analysis module calculates and outputs the structural convergence tension index (STCI) based on standardized vitiligo data, sets the convergence interval threshold for preliminary comparative evaluation, and triggers skin thermal recovery analysis based on the preliminary comparative evaluation results.

[0065] The thermal recovery analysis module extracts thermal recovery features from infrared white spot area images to obtain standardized thermal recovery data, and then calculates and outputs the thermal response consistency mapping index HRCI based on the standardized thermal recovery data;

[0066] The comprehensive analysis module summarizes and calculates the thermal response consistency mapping index HRCI and the structural convergence tension index STCI, outputs the comprehensive matching index SHMI, and sets the recovery interval threshold and the comprehensive matching index SHMI for secondary comparative evaluation.

[0067] In this embodiment, the system provides an intelligent management system for chronic diseases that integrates visual perception, AI image processing, and multi-source indicator fusion evaluation, and is particularly suitable for the management of skin diseases such as vitiligo that require continuous monitoring of treatment responses. The laser response capture module collects RGB images and infrared thermal map data after laser treatment in real time, and combines the multidimensional data processing module to standardize, semantically segment, and extract features of the images. The system can accurately obtain the structural parameters and thermal parameters of the white spot area, and use the structural convergence analysis module to calculate the structural convergence tension index STCI to quantify the boundary change trend. Further, under the triggering of a specific threshold, the thermal recovery analysis module calculates the thermal response consistency mapping index HRCI to reflect the regional metabolic activity and rewarming stability. Finally, the comprehensive analysis module generates the structural-thermal resonance matching index SHMI by summarizing the thermal response consistency mapping index HRCI and the structural convergence tension index STCI, and performs response level division and treatment strategy feedback based on the preset level threshold, thereby realizing closed-loop control and personalized adjustment of the laser treatment response. Through the implementation of this system, the joint perception and dynamic analysis of the morphological evolution and thermal metabolic behavior of vitiligo areas has been achieved, overcoming the traditional reliance on manual visual inspection, the highly subjective response assessment, and the lack of quantification. It also introduced a structural and thermal resonance index system for the first time, establishing a coupled mapping mechanism between morphological changes and metabolic trends, improving the objectivity, accuracy, and timeliness of vitiligo laser treatment response judgments. The system boasts fully automatic, multi-dimensional, closed-loop, and quantifiable management features, making it suitable not only for tracking the efficacy of vitiligo patients and regulating treatment strategies, but also for promoting the intelligent assessment system for chronic diseases in the management of other skin and metabolic diseases. It has significant clinical application value and promotion prospects.

[0068] Example 2

[0069] See also Figure 1 ,Specifically: the laser response capture module includes a white spot image capture unit and an image data transmission unit;

[0070] The vitiligo image capture unit uses a visual sensor to collect RGB image data sets and infrared thermal map data sets of the skin area after laser treatment in real time by setting the image collection time points after laser treatment.

[0071] The image acquisition time points were set at 12 h, 48 h, and 96 h after each laser treatment;

[0072] Vision sensors include RGB vision sensors and thermal imaging vision sensors;

[0073] The image data transmission unit uses the built-in wireless communication module of the visual sensor and sets up a 5G communication network to wirelessly connect the visual sensor with the information management system, and transmits the RGB image data set and infrared thermal map data set to the information management system in real time.

[0074] In this embodiment, the laser response capture module of the system realizes multimodal, time-series image acquisition and efficient transmission of the white spot area after laser treatment by constructing a linkage mechanism between the white spot image capture unit and the image data transmission unit. In the white spot image capture unit, by presetting the image acquisition time points, namely 12 hours, 48 ​​hours and 96 hours after laser treatment, the visual sensor with integrated RGB and infrared functions is used to synchronously obtain the visible light structural information and skin surface thermal distribution data of the white spot area, laying a data foundation for subsequent structural changes and thermal response analysis. At the same time, the image data transmission unit is connected to the 5G communication network through the built-in wireless communication module of the visual sensor, realizing the real-time upload and remote transmission of high-frequency, high-resolution image data, effectively avoiding the traditional data collection method that relies on manual export and has high latency. Through the above implementation mode, this module realizes the standardization, timeliness and high-precision sampling of image acquisition after white spot treatment, and significantly improves the system's recording granularity and coverage density of the entire white spot evolution process. At the same time, with the help of high-speed 5G networks and edge device intelligent modules, this system supports remote dynamic intervention and intelligent monitoring, which not only shortens the doctor's judgment delay, but also provides continuous and traceable training data for the AI ​​model, greatly improving the real-time, accuracy and automation of subsequent data analysis and treatment response evaluation.

[0075] Example 3

[0076] See also Figure 1 and Figure 3 ,Specifically: the multi-dimensional data processing module includes an AI embedding unit and an RGB image processing unit;

[0077] The AI ​​embedding unit embeds an AI image processing model in the information management system and inputs the real-time received RGB image dataset and infrared thermal image dataset into the AI ​​image processing model for preprocessing;

[0078] The AI ​​image processing model includes an image normalization preprocessing module, an RGB feature extraction module, and a thermal recovery feature extraction module;

[0079] The image standardization preprocessing module loads OpenCV image processing technology into the AI ​​image processing model to standardize the RGB image dataset and infrared thermal map dataset, and performs size unification, image alignment, intensity normalization, and color correction on the RGB image dataset and infrared thermal map dataset;

[0080] The white spot areas in all images in the RGB image dataset and infrared thermal map dataset after image standardization preprocessing are outlined using medical image annotation software, and an AI segmentation model is constructed using the semantic segmentation network U-Net. A large number of samples with the white spot area outlines are input into the AI ​​segmentation model to train the AI ​​segmentation model, and the real-time RGB image dataset and infrared thermal map dataset are input into the AI ​​segmentation model to segment the RGB image dataset and the infrared thermal map dataset, and extract the RGB white spot area images and the infrared white spot area images.

[0081] The RGB image processing unit obtains the RGB white spot area image through the AI ​​image processing model and transmits it to the RGB feature extraction module to perform feature extraction to obtain white spot morphological data;

[0082] The white spot morphological data include the white spot area Sb(t) at time t, the edge curvature distribution kurtosis Pc(t) at time t, and the tangential direction disturbance amplitude Ttan(t) at time t;

[0083] The white spot morphology data were normalized using Z-score to eliminate the dimensions of all parameters in the white spot morphology data and unify the units;

[0084] The white spot area Sb(t) at time t is calculated and output by extracting the RGB white spot area image and summing up the pixel points of the white spot area;

[0085] The edge curvature distribution kurtosis Pc(t) at time t is obtained by using OpenCV to obtain the contour curve of the whiteboard area. At the same time, the GCN graph convolutional neural network is introduced into the RGB feature extraction module to extract the edge curvature distribution kurtosis Pc(t) at time t of the contour curve;

[0086] The tangent direction disturbance amplitude Ttan(t) at time t is obtained by embedding time-series image contour point tracking and tangent vector cosine similarity in the RGB feature extraction module. Through time-series image contour point tracking, the matching points between the two RGB white spot area images at different acquisition time points are extracted based on the tangent vector cosine similarity, and the similarity between the two RGB white spot area images is calculated to obtain the tangent direction disturbance amplitude Ttan(t) at time t. The higher the disturbance amplitude, the lower the similarity, indicating that the tangent of the white spot boundary changes greatly, and the morphology has undergone obvious changes such as contraction, expansion, and rotation. The lower the disturbance amplitude, the higher the similarity, indicating that the local direction of the two white spot contours changes little and the boundary shape difference is small.

[0087] In this embodiment, the multi-dimensional data processing module of the system realizes a complete chain-type processing flow of standardized preprocessing, regional segmentation, feature extraction and parameter quantification of white spot area image data by constructing a collaborative working mechanism between the AI ​​embedding unit and the RGB image processing unit. The AI ​​embedding unit synchronously inputs the collected RGB and infrared thermal image data into the AI ​​image processing model that integrates model technologies such as OpenCV, U-Net, GCN, etc. First, the input image is resized, aligned, brightness normalized and color corrected through the image standardization preprocessing module, solving the problem of inconsistent image quality across devices and time. Subsequently, the system uses the U-Net semantic segmentation network trained with doctor-labeled samples to realize automatic recognition and segmentation of white spot areas, greatly reducing the cost and error of manual intervention. Furthermore, the RGB image processing unit calls the segmented white spot image data to extract key structural parameters: white spot area Sb, edge curvature kurtosis Pc and tangent direction perturbation amplitude Ttan, and by introducing the GCN graph neural network, temporal contour tracking and cosine similarity algorithm, it accurately depicts the geometric evolution trend and boundary dynamic change characteristics of the white spot area over time. All feature parameters are uniformly standardized using the Z-score before processing to eliminate dimensional interference and ensure the computational stability and generalization ability of the subsequent algorithm model. In summary, this module achieves efficient conversion from raw images to structural parameters during implementation, which not only improves the automation and accuracy of vitiligo image analysis, but also provides multi-dimensional, quantitative and traceable core input data for subsequent structural trend analysis and thermal coupling evaluation. Ultimately, this module significantly improves the system's recognition accuracy of the dynamic recovery of the structure of the lesion area, the consistency of parameter interpretation, and the model's support for efficacy judgment, laying a solid foundation for building a quantifiable and feedback-based personalized chronic disease management system.

[0088] Example 4

[0089] See also Figure 1 ,Specifically: the structural convergence analysis module includes the white spot structure convergence ,analysis unit and the convergence evaluation unit;

[0090] The white spot structure convergence analysis unit extracts the white spot morphology data, calculates and outputs the structural convergence tension index STCI, and measures the convergence of the white spot boundary;

[0091] The structural convergence tension index STCI is calculated and output by the following algorithm formula;

[0092] ;

[0093] Where n represents the total number of image acquisition time points. Here, the image acquisition time points are 12 hours, 48 ​​hours, and 96 hours, so the total number of image acquisition time points n=3. cos represents the cosine function. represents the standard deviation of the curvature distribution, d represents the calculus function, dt represents the time calculus function, represents the directional similarity factor, t represents the time, that is, the time of image acquisition. Specifically, t=1 represents the 12th hour after laser treatment, t=2 represents the 48th hour after laser treatment, and t=3 represents the 96th hour after laser treatment.

[0094] This formula is a product-type composite average function used to quantify the structural change trend of the white spot area after laser treatment of vitiligo. It is specifically divided into three items:

[0095] Indicates the area change rate, indicating whether the morphology is in a "convergence" trend. The absolute value is used to accommodate expansion and contraction. A larger value indicates a significant area change and a positive response.

[0096] It represents the curvature standard deviation suppression term. The more irregular the curvature and the larger the standard deviation, the smaller the term is, indicating an unstable structure and a large penalty coefficient, which emphasizes the importance of “neat boundaries” for the judgment of white spot convergence.

[0097] It represents the directional consistency term. If the boundary tangent direction is basically consistent, the value is close to 0 if the boundary direction is severely disturbed.

[0098] The user aggregates the three image acquisition time points for mean calculation.

[0099] The convergence assessment unit sets the convergence interval threshold, which includes the first convergence interval threshold F1 and the second convergence interval threshold F2. It then performs a preliminary comparative assessment of the real-time acquired structural convergence tension index STCI with the convergence interval threshold to determine the structural recovery after vitiligo laser treatment. Based on the preliminary comparative assessment results, it triggers skin thermal recovery analysis. The specific assessment contents are as follows;

[0100] The structural convergence tension index (STCI) of historical laser treatment effective and ineffective samples was extracted, and the average value of the concentrated segment of 85% effective samples was set as the first convergence interval threshold F1, and the average value of the concentrated segment of 90% ineffective samples was set as the second convergence interval threshold F2;

[0101] When the structural convergence tension index STCI is greater than the first convergence interval threshold F1, it indicates that the white spot structure is stably converging, that is, the white spot area continues to shrink, the boundary is smooth and stable, and the shrinkage direction is consistent, and no adjustment is performed at this time;

[0102] When the second convergence interval threshold F2 < structural convergence tension index STCI ≤ the first convergence interval threshold F1, it indicates that there is a fluctuation response in the white spot structure contraction, and at this time, it is prompted to perform image acquisition and analysis again, that is, the white spot image capture unit is performed again at 12 hours, 48 ​​hours and 96 hours after 96 hours;

[0103] When the structural convergence tension index STCI ≥ the second convergence interval threshold F2, it indicates that the white spot structure is unresponsive, and the skin thermal recovery analysis is triggered.

[0104] In this embodiment, in the structural convergence analysis module of the system, a set of intelligent analysis mechanisms for quantifying the structural change trend of the white spot area and dynamically distinguishing the therapeutic response status is constructed through the collaborative integration of the white spot structure convergence analysis unit and the convergence evaluation unit. In terms of specific implementation methods, the system first extracts the white spot morphological data generated by the multidimensional data processing module, and based on the image acquisition results of three different time points, applies the product-type composite average function to calculate the output structural convergence tension index STCI. This index comprehensively reflects whether the white spot area changes stably, whether the boundary is smooth and regular, and whether the morphological changes are consistent. It has high accuracy and clear physical meaning. In order to ensure that the evaluation results have clinical decision-making value and data-driven logic support, the system also sets a convergence interval threshold derived from historical samples, which is used to distinguish the three types of recovery states: "structural stable convergence", "structural fluctuation response" and "structural non-response", and automatically triggers whether to enter the thermal recovery analysis stage or whether to re-acquire images based on the corresponding state, thereby constructing an adaptive feedback closed-loop mechanism. Among them, the first convergence interval threshold F1 is determined based on the STCI mean of 85% of the effective laser samples, and the second convergence interval threshold F2 is set based on 90% of the invalid samples, ensuring that the interval division has a high degree of discrimination and medical interpretability. In summary, the implementation of this module effectively realizes the digital modeling and intelligent judgment of the structural changes of white spots after laser treatment of vitiligo, greatly improving the doctor's ability to objectively evaluate the trend of structural changes and the efficiency of response adjustment. It not only reduces the treatment misunderstandings caused by subjective judgment deviations, but also provides a clear and scientific decision-making precondition for the subsequent system thermal recovery analysis, and ultimately promotes the precise regulation of the treatment path and closed-loop tracking of efficacy, which is a key link in the intelligent process of vitiligo treatment.

[0105] Example 5

[0106] See also Figure 1 and Figure 2 ,Specifically: the thermal recovery analysis module includes a thermal recovery ,feature extraction unit and a thermal response consistency analysis unit;

[0107] After triggering the skin thermal recovery analysis, the thermal recovery feature extraction unit extracts the infrared white spot area image processed by the image standardization preprocessing module and inputs it into the thermal recovery feature extraction module in the AI ​​image processing model. The thermal recovery feature of the infrared white spot area image is extracted, and the feature extraction result is normalized using Z-score to eliminate the dimensional influence of all extraction results to obtain standardized thermal recovery data.

[0108] The standardized thermal recovery data include the temperature gradient recovery value △Tr(t) at time t, the local curvature of the temperature recovery curve Kf(t) at time t, and the regional heat flow reconstruction density Rs(t) at time t;

[0109] The temperature gradient recovery value △Tr(t) at time t is obtained by performing difference calculation between the average temperature of the white spot area in the infrared white spot area image and the average temperature of the surrounding skin in the three image acquisition time points through the thermal recovery feature extraction module;

[0110] The local curvature Kf(t) of the temperature recovery curve at time t is obtained by fitting the temperature of the white spot area within the three image acquisition time points, and calculating the second-order derivative of the temperature curve at the current time point;

[0111] The regional heat flow reconstruction density Rs(t) at time t is obtained by embedding a thermal map regression network in the thermal recovery feature extraction module to invert the heat flux vector of each pixel point in the white spot area of ​​all infrared white spot area images, and integrating and averaging the heat vector of each pixel point in the white spot area.

[0112] The thermal response consistency analysis unit extracts standardized thermal recovery data, calculates and outputs the thermal response consistency mapping index HRCI, and analyzes the coupling strength between thermal response and structural trend;

[0113] The thermal response consistency mapping index HRCI is calculated and output by the following algorithm formula;

[0114] ;

[0115] Where HRCI(t) represents the thermal response consistency mapping index at time t, T represents the image acquisition time interval, such as 96 hours after laser treatment, log represents the logarithmic function, d represents the calculus function, and dt represents the time calculus function;

[0116] Indicates the rate of change of temperature difference. If the temperature difference decreases rapidly over time, it means that the temperature of the white spot area is returning to normal body temperature, indicating that blood flow and energy metabolism are recovering;

[0117] Indicates the temperature recovery stability modulation factor. If the temperature change curvature is large, it means the recovery is unstable, with "fallback" and "jitter" phenomena, which is bad. At the same time, if the curvature is 0, it means extremely smooth recovery, which is good. Therefore, this item suppresses the unstable area.

[0118] Indicates the intensity of heat flow metabolic activity and is used to analyze whether there is persistent heat diffusion in the vitiligo area, which is usually caused by metabolism or microcirculation. The logarithmic function log form can prevent abnormally high heat flow values ​​from distorting the overall results and increase system stability.

[0119] The significance of the three products lies in multiplying rapid recovery, stable trend and active metabolism, using integral accumulation to evaluate the consistency of metabolic and structural responses in the entire efficacy analysis window, such as 0–96 hours, and analyzing the coupling strength between thermal response and structural trend.

[0120] In this embodiment, the system, through the collaborative design of a thermal recovery feature extraction unit and a thermal response consistency analysis unit, achieves intelligent analysis of the thermal metabolic changes in the vitiligo area and coupled matching assessment of the structural response. Specifically, the system first invokes the thermal recovery analysis process through a pre-trigger mechanism. The infrared image of the vitiligo area, pre-processed by the image normalization module, is input into the thermal recovery feature extraction module of the AI ​​image processing model. This module utilizes a deep learning thermal map regression network combined with Z-score normalization technology to accurately output standardized thermal recovery data corresponding to three key indicators: the rate of recovery of surface temperature difference, the stability of the recovery process, and the level of tissue metabolic activity. Subsequently, the thermal response consistency analysis unit incorporates an integral coupled discriminant model, integrating these three core parameters to calculate and output the thermal response consistency mapping index (HRCI). This index has clear physical meaning and discernible power: it measures the temperature difference recovery rate to determine the degree of blood flow and metabolic recovery in the vitiligo area. It also utilizes the interference of local temperature curvature modulation instability on the therapeutic effect, and combines it with heat flux density to assess continuous thermal diffusion behavior, thus forming a comprehensive assessment system that integrates structure and metabolism. Through the implementation of this module, not only has quantitative modeling and trend analysis been achieved for the local metabolic response of the skin after laser treatment for vitiligo been achieved, but it has also effectively established quantitative indicators for the temporal synchronization and spatial consistency between structural changes and thermal responses, providing a critical physiological basis for subsequent comprehensive efficacy analysis and treatment plan feedback. Overall, this module significantly improves the accuracy of the chronic disease digital management system's perception of the lesion recovery process and the intelligent level of efficacy response judgment, serving as a core support link in promoting the transition from subjective experience-based diagnosis to visual quantitative diagnosis and treatment.

[0121] Example 6

[0122] See also Figure 1,Specifically: the comprehensive analysis module includes a structural and ,thermal resonance analysis unit and a comprehensive evaluation unit;

[0123] The structural and thermal resonance analysis unit calculates and outputs the comprehensive matching index SHMI based on the obtained thermal response consistency mapping index HRCI and structural convergence tension index STCI, which measures the current morphological convergence trend and thermal metabolic coupling behavior of vitiligo.

[0124] The comprehensive matching index SHMI is calculated and output by the following algorithm formula;

[0125] ;

[0126] Where SHMI(t) represents the comprehensive matching index at time t, represents the average value of the local curvature of the temperature recovery curve;

[0127] Represents the resonance consistency product, indicating whether the structural recovery trend and thermal metabolic response are significant at the same time. Only when the structural change is obvious and the thermal flow response is consistent will the product value be large, indicating resonance recovery and is a positive performance factor for recovery.

[0128] Indicates the mismatch between structure and thermal performance. If the structure changes rapidly but the temperature fluctuates greatly, , or if the trends of the two are inconsistent, it means that the metabolism has not kept up with the structural changes and there is a mismatch. The purpose of adding 1 is to prevent the denominator from approaching 0 and causing numerical divergence, to ensure the stability of the system, and it is an overall stable adjustment item.

[0129] The comprehensive evaluation unit sets a recovery interval threshold and conducts a secondary comparative evaluation with the comprehensive matching index SHMI obtained in real time to judge the recovery status of the current white spot area, and divides the recovery level based on the secondary comparative evaluation results, and performs corresponding information management based on the divided recovery level;

[0130] The recovery interval threshold includes a first recovery interval threshold Q1 and a second recovery interval threshold Q2;

[0131] The specific assessment contents are as follows;

[0132] When the comprehensive matching index SHMI is greater than the first recovery interval threshold Q1, it is classified as recovery level A, that is, significant recovery. At this time, the current governance plan is maintained without intervention;

[0133] When the second recovery interval threshold Q2 is less than the comprehensive matching index SHMI ≤ the first recovery interval threshold Q1, it is classified as recovery level B, which means that there is recovery, but there is still a slight mismatch in structure or heat. At this time, it is suggested to increase the frequency of laser treatment by 50%, such as from every 4 days to every 2 days. At the same time, a melanin nutrition reminder is sent to remind you to supplement tyrosinase;

[0134] When the comprehensive matching index SHMI ≤ the second recovery interval threshold Q2, it is classified as recovery level C, that is, the structure and thermal response are weak or inconsistent, and the therapeutic effect is low. At this time, a re-examination prompt for the current white spot area is generated, and the laser parameters are doubled;

[0135] Among them, the first recovery interval threshold Q1 and the second recovery interval threshold are obtained by using clinical samples in the early data training stage, in which each sample has a manually calibrated recovery level label, and calculating the comprehensive matching index SHMI of all valid samples. It is found that: the mean of the comprehensive matching index SHM concentrated segment of 80% significantly recovered patients is set as the first recovery interval threshold Q1, and the mean of the comprehensive matching index SHMI concentrated segment of 90% invalid or insignificant recovery is set as the second recovery interval threshold Q2.

[0136] In this embodiment, the method realizes highly integrated discrimination and intelligent feedback of the treatment effect of vitiligo by integrating the structural and thermal resonance analysis unit and the comprehensive evaluation unit. In terms of the specific implementation method, the system takes the structural convergence tension index STCI and the thermal response consistency mapping index HRCI obtained in the previous module as input features, and submits them to the structural and thermal resonance analysis unit for fusion analysis. Based on the function model of the resonance consistency product and the structural-thermal mismatch, the comprehensive matching index SHMI representing the resonance level of the therapeutic effect is calculated and output. This index not only quantifies the degree of consistency between the structural morphological changes of vitiligo and the local thermal metabolic response, but also suppresses data mutations and abnormal fluctuations through robust adjustment items, thereby ensuring the robustness and stability of the efficacy evaluation. On this basis, the comprehensive evaluation unit conducts a secondary comparative evaluation of the comprehensive matching index SHMI based on the recovery interval threshold extracted by clinical sample data training, so as to realize the classification of the current treatment response of individual patients and the matching adjustment of treatment strategies. The beneficial effects brought about by the implementation of this module are mainly reflected in three aspects: first, it breaks the subjective judgment barrier in traditional vitiligo treatment and builds an objective and quantifiable efficacy feedback loop; second, it realizes the unified evaluation of structural changes and metabolic responses, and enhances the scientific nature of the identification of treatment response mechanisms; third, it provides a data-driven personalized treatment regulation mechanism, improves the efficiency of efficacy improvement and patient compliance, and significantly promotes the refinement, automation and precision of intelligent diagnosis and treatment management of chronic diseases.

[0137] While the embodiments of the present invention have been shown and described, it will be apparent to those skilled in the art that various changes, modifications, substitutions, and alterations can be made to the embodiments without departing from the principles and spirit of the invention.

Claims

1. An AI-based intelligent management system for multi-disease chronic disease information, characterized by: It includes laser response capture module, multi-dimensional data processing module, structural convergence analysis module, thermal recovery analysis module and comprehensive analysis module; The laser response capture module collects the RGB image dataset and infrared thermal map dataset of the skin area after laser treatment in real time by using a visual sensor, and transmits the RGB image dataset and infrared thermal map dataset to the information management system; The multidimensional data processing module preprocesses the RGB image dataset and the infrared thermal image dataset by using an AI image processing model to obtain an RGB white spot area image and an infrared white spot area image, and extracts features from the RGB white spot area image to obtain standardized white spot data; The multi-dimensional data processing module includes an RGB image processing unit; The RGB image processing unit obtains the RGB white spot area image through the AI ​​image processing model and transmits it to the RGB feature extraction module to perform feature extraction to obtain white spot morphological data; The white spot morphological data includes the white spot area Sb(t) at time t, the edge curvature distribution kurtosis Pc(t) at time t, and the tangential direction disturbance amplitude Ttan(t) at time t; The white spot morphology data were normalized using Z-score to eliminate the dimensions of all parameters in the white spot morphology data and unify the units; The structural convergence analysis module calculates and outputs the structural convergence tension index (STCI) based on the standardized vitiligo data, sets a convergence interval threshold for preliminary comparative evaluation, and triggers skin thermal recovery analysis based on the preliminary comparative evaluation results; The structural convergence analysis module includes a white spot structure convergence analysis unit and a convergence evaluation unit; The white spot structure convergence analysis unit extracts white spot morphological data and calculates and outputs a structural convergence tension index STCI to measure the convergence of the white spot boundary; The structural convergence tension index STCI is calculated and output by the following algorithm formula: ； Where n represents the total number of image acquisition time points, cos represents the cosine function, represents the standard deviation of the curvature distribution, d represents the calculus function, dt represents the time calculus function, represents the direction similarity factor, t represents the time; The thermal recovery analysis module extracts thermal recovery features from the infrared white spot area image to obtain standardized thermal recovery data, and calculates and outputs the thermal response consistency mapping index HRCI based on the standardized thermal recovery data; The comprehensive analysis module summarizes and calculates the thermal response consistency mapping index HRCI and the structural convergence tension index STCI, outputs the comprehensive matching index SHMI, and sets the recovery interval threshold and the comprehensive matching index SHMI for secondary comparative evaluation.

2. The AI-based intelligent management system for multiple chronic disease information according to claim 1 is characterized by: The laser response capture module includes a white spot image capture unit and an image data transmission unit; The vitiligo image capture unit uses a visual sensor to collect RGB image data sets and infrared thermal image data sets of the skin area after laser treatment in real time by setting image collection time points after laser treatment. The image acquisition time points were set at 12 hours, 48 ​​hours, and 96 hours after each laser treatment; The visual sensor includes an RGB visual sensor and a thermal imaging visual sensor; The image data transmission unit uses the built-in wireless communication module of the visual sensor and sets up a 5G communication network to wirelessly connect the visual sensor with the information management system, and transmits the RGB image data set and the infrared thermal map data set to the information management system in real time.

3. The AI-based intelligent management system for multiple chronic disease information according to claim 2 is characterized by: The multi-dimensional data processing module also includes an AI embedding unit and; The AI ​​embedding unit embeds an AI image processing model in the information management system and inputs the real-time received RGB image dataset and infrared thermal image dataset into the AI ​​image processing model for preprocessing; The AI ​​image processing model includes an image standardization preprocessing module, an RGB feature extraction module, and a thermal recovery feature extraction module; The image standardization preprocessing module loads OpenCV image processing technology into the AI ​​image processing model to standardize the RGB image dataset and the infrared thermal map dataset, and performs size unification, image alignment, intensity normalization and color correction on the RGB image dataset and the infrared thermal map dataset; The white spot areas in all images in the RGB image dataset and infrared thermal map dataset after image standardization preprocessing are outlined using medical image annotation software, and an AI segmentation model is constructed using the semantic segmentation network U-Net. A large number of samples with the white spot area outlines are input into the AI ​​segmentation model to train the AI ​​segmentation model, and the real-time RGB image dataset and infrared thermal map dataset are input into the AI ​​segmentation model to segment the RGB image dataset and the infrared thermal map dataset, and extract the RGB white spot area images and the infrared white spot area images.

4. The AI-based intelligent management system for multiple chronic disease information according to claim 1 is characterized by: The convergence evaluation unit sets a convergence interval threshold, which includes a first convergence interval threshold F1 and a second convergence interval threshold F2. The convergence evaluation unit performs a preliminary comparative evaluation of the structural convergence tension index STCI obtained in real time with the convergence interval threshold to judge the recovery of the structure after the vitiligo laser treatment, and triggers the skin thermal recovery analysis based on the preliminary comparative evaluation results. The specific evaluation contents are as follows; When the structural convergence tension index STCI is greater than the first convergence interval threshold F1, it indicates that the white spot structure is stably converging, that is, the white spot area continues to shrink, the boundary is smooth and stable, and the shrinkage direction is consistent, and no adjustment is performed at this time; When the second convergence interval threshold F2 < structural convergence tension index STCI ≤ the first convergence interval threshold F1, it indicates that there is a fluctuation response in the white spot structure contraction, and it is prompted to execute the image acquisition time point again for acquisition and analysis; When the structural convergence tension index STCI ≥ the second convergence interval threshold F2, it indicates that the white spot structure is unresponsive, and the skin thermal recovery analysis is triggered.

5. The AI-based intelligent management system for multiple chronic disease information according to claim 1 is characterized by: The thermal recovery analysis module includes a thermal recovery feature extraction unit and a thermal response consistency analysis unit; The thermal recovery feature extraction unit extracts the infrared white spot area image processed by the image standardization preprocessing module after triggering the skin thermal recovery analysis, inputs the image into the thermal recovery feature extraction module in the AI ​​image processing model, extracts the thermal recovery features of the infrared white spot area image, and performs Z-score normalization on the feature extraction results to eliminate the dimensional influence of all extraction results, thereby obtaining standardized thermal recovery data; The standardized thermal recovery data includes the temperature gradient recovery value ΔTr(t) at time t, the local curvature Kf(t) of the temperature recovery curve at time t, and the regional heat flux reconstruction density Rs(t) at time t.

6. The AI-based intelligent management system for multiple chronic disease information according to claim 5 is characterized by: The thermal response consistency analysis unit extracts standardized thermal recovery data, calculates and outputs a thermal response consistency mapping index HRCI, and analyzes the coupling strength between thermal response and structural trend; The thermal response consistency mapping index HRCI is calculated and output by the following algorithm formula: ； Where HRCI(t) represents the thermal response consistency mapping index at time t, T represents the image acquisition time interval, log represents the logarithmic function, d represents the calculus function, and dt represents the time calculus function.

7. The AI-based intelligent management system for multiple chronic disease information according to claim 6 is characterized by: The comprehensive analysis module includes a structural and thermal resonance analysis unit and a comprehensive evaluation unit; The structural and thermal resonance analysis unit calculates and outputs a comprehensive matching index SHMI based on the acquired thermal response consistency mapping index HRCI and structural convergence tension index STCI, thereby measuring the morphological convergence trend and thermal metabolic coupling behavior of the current vitiligo. The comprehensive matching index SHMI is calculated and output by the following algorithm formula: ； Where SHMI(t) represents the comprehensive matching index at time t, Represents the average value of the local curvature of the temperature recovery curve.

8. The AI-based intelligent management system for multiple chronic disease information according to claim 7 is characterized by: The comprehensive evaluation unit performs a secondary comparative evaluation on the recovery interval threshold and the comprehensive matching index SHMI obtained in real time, judges the recovery status of the current white spot area, and divides the recovery level based on the secondary comparative evaluation result, and performs corresponding information management based on the divided recovery level; The recovery interval threshold includes a first recovery interval threshold Q1 and a second recovery interval threshold Q2; The specific assessment contents are as follows; When the comprehensive matching index SHMI is greater than the first recovery interval threshold Q1, it is classified as recovery level A. At this time, the current governance plan is maintained without intervention; When the second recovery interval threshold Q2 is less than the comprehensive matching index SHMI ≤ the first recovery interval threshold Q1, it is classified as recovery level B. At this time, it is prompted to increase the frequency of laser treatment by 50%, and a melanin nutrition reminder is sent to remind you to supplement tyrosinase; When the comprehensive matching index SHMI ≤ the second recovery interval threshold Q2, it is classified as recovery level C. At this time, a re-examination prompt for the current white spot area is generated, and the laser parameters are doubled.

Images