Visual data processing system applied to pediatric department of traditional Chinese medicine

Abstract

The invention relates to the technical field of data management, in particular to a visual data processing system applied to traditional Chinese medicine pediatrics. The system comprises the following modules: a data acquisition module used for acquiring a pediatric physiological time sequence data set; carrying out physiological event anchor point identification on the pediatric physiological time sequence data set to obtain a physiological event anchor point set; the time calibration module is used for performing noise suppression on the pediatric physiological time sequence data set according to the physiological event anchor point set to obtain a noise-suppressed physiological time sequence data set; performing time calibration on the noise suppression physiological time sequence data set to obtain a time correction physiological data set; the feature extraction module is used for extracting a tongue picture image set from the time correction physiological data set; and generating a tongue picture feature vector set based on the tongue picture image set. According to the method, the problem of dynamic association analysis caused by multi-dimensional sign data isomerism in traditional Chinese medicine pediatric clinical data processing can be effectively solved.

Description

Technical Field

[0001] The present invention relates to the technical field of data management, and in particular to a visualization data processing system applied to traditional Chinese medicine pediatrics. Background Art

[0002] In the early stage, the data processing of traditional Chinese medicine pediatrics mainly relied on manual records and simple text descriptions, making it difficult to systematically manage and deeply analyze complex physiological sign data. With the progress of medical device technology, multi-modal detection technologies such as tongue coating microscopic imaging, pulse waveform detection, and body surface thermal imaging have gradually been applied to the clinical practice of traditional Chinese medicine pediatrics, providing rich data sources for traditional Chinese medicine. However, some problems that need to be solved urgently have also emerged during the development of these technologies.

[0003] When dealing with multi-dimensional sign data, existing traditional Chinese medicine pediatrics data processing systems often have difficulty in achieving effective dynamic correlation analysis due to data heterogeneity. Traditional visualization systems usually can only display different modal data such as tongue coating microscopic images, pulse waveform signals, and constitution identification scales in isolation, lacking the ability to construct a time-series correlation composite view. For example, in the tracking scenario of childhood allergic purpura, it is necessary to comprehensively observe the body surface thermal imaging distribution of purpura on the limbs of children, the microcirculation state of the sublingual collaterals, and the change trend of the degree of slippery and rapid pulse. However, due to the time stamp deviation of different detection devices, the system cannot automatically align the acquisition time nodes of multi-source data, resulting in easy time-domain misalignment judgment when comparing the evolution of tongue image features and the fluctuation of pulse parameters, thus affecting the accurate grasp of the development process of children's conditions. Summary of the Invention

[0004] Based on this, it is necessary for the present invention to provide a visualization data processing system applied to traditional Chinese medicine pediatrics to solve at least one of the above technical problems.

[0005] To achieve the above object, a visualization data processing system applied to traditional Chinese medicine pediatrics includes the following modules: A data acquisition module, configured to acquire a pediatric physiological time-series data set; identify physiological event anchor points for the pediatric physiological time-series data set to obtain a physiological event anchor point set; A time calibration module, configured to perform noise suppression on the pediatric physiological time-series data set according to the physiological event anchor point set to obtain a noise-suppressed physiological time-series data set; perform time calibration on the noise-suppressed physiological time-series data set to obtain a time-corrected physiological data set; A feature extraction module, configured to extract a tongue image set from the time-corrected physiological data set; generate a tongue image feature vector set based on the tongue image set; extract a pulse signal data set from the time-corrected physiological data set; construct a pulse characteristic parameter set based on the pulse signal data set; extract a thermal imaging image sequence from the time-corrected physiological data set; generate a body area temperature feature table based on the thermal imaging image sequence; An evolution index construction module, configured to construct a tongue image time series index sequence based on a tongue image feature vector set; generate a pulse image time series index sequence based on a pulse characteristic parameter set; generate a body surface temperature rhythm index sequence based on a body region temperature feature table; A visualization module, configured to construct a visual coding rule for pediatric data based on the tongue image time series index sequence, the pulse image time series index sequence, and the body surface temperature rhythm index sequence; visualize the tongue image time series index sequence, the pulse image time series index sequence, and the body surface temperature rhythm index sequence according to the visual coding rule for pediatric data.

[0006] In the present invention, through the data acquisition module, a comprehensive acquisition of the pediatric physiological time series data set is realized, and through the identification of physiological event anchor points, a key time reference point is provided for subsequent processing. Through the time calibration module, noise suppression and time calibration are performed based on these anchor points, thereby eliminating the problem of time stamp deviation between different detection devices, and ensuring the consistency and accuracy of multi-source data in the time dimension. Through the feature extraction module, key feature information of tongue images, pulse images, and thermal imaging can be extracted from the calibrated data, and converted into feature vectors or parameter sets for analysis. Through the evolution index construction module, these feature information are further converted into time series index sequences, capturing the dynamic change law of pediatric physiological data over time. Through the visualization module, a visual coding rule for pediatric data is constructed based on these time series index sequences, and an intuitive visualization display is performed, enabling the evolution process of tongue image, pulse image, and body surface temperature data to be clearly observed. In summary, the present invention can effectively solve the problem of dynamic correlation analysis caused by the heterogeneity of multi-dimensional physical sign data in the clinical data processing of traditional Chinese medicine pediatrics. Description of the Drawings

[0007] By reading the following detailed description with reference to the accompanying drawings, other features, objects, and advantages of the present invention will become more apparent: Figure 1 Shows a schematic diagram of the module process of an application to a visual data processing system for traditional Chinese medicine pediatrics in an embodiment. Detailed Embodiments

[0008] The technical method of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative efforts shall fall within the scope of protection of the present invention.

[0009] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and thus repeated descriptions thereof will be omitted. Some of the block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities may be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor methods and / or microcontroller methods.

[0010] It should be understood that although terms such as "first" and "second" may be used herein to describe various units, these units should not be limited by these terms. These terms are only used to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0011] To achieve the above object, please refer to Figure 1 , the present invention provides a visualization data processing system applied to traditional Chinese medicine pediatrics, including the following modules: S1: A data acquisition module, configured to acquire a pediatric physiological time series dataset; perform physiological event anchor point recognition on the pediatric physiological time series dataset to obtain a physiological event anchor point set; S2: A time calibration module, configured to perform noise suppression on the pediatric physiological time series dataset according to the physiological event anchor point set to obtain a noise-suppressed physiological time series dataset; perform time calibration on the noise-suppressed physiological time series dataset to obtain a time-corrected physiological dataset; S3: A feature extraction module, configured to extract a tongue image set from the time-corrected physiological dataset; generate a tongue image feature vector set based on the tongue image set; extract a pulse signal dataset from the time-corrected physiological dataset; construct a pulse characteristic parameter set based on the pulse signal dataset; extract a thermal imaging image sequence from the time-corrected physiological dataset; generate a body area temperature feature table based on the thermal imaging image sequence; S4: An evolution index construction module, configured to construct a tongue image time series index sequence based on the tongue image feature vector set; generate a pulse time series index sequence based on the pulse characteristic parameter set; generate a body surface temperature rhythm index sequence based on the body area temperature feature table; S5: A visualization module, configured to construct a visual coding rule for pediatric data based on the tongue image time series index sequence, the pulse time series index sequence, and the body surface temperature rhythm index sequence; perform visualization on the tongue image time series index sequence, the pulse time series index sequence, and the body surface temperature rhythm index sequence according to the visual coding rule for pediatric data.

[0012] In the present invention, the comprehensive collection of pediatric physiological time series data sets is achieved through the data acquisition module, and through the identification of physiological event anchor points, key time reference points are provided for subsequent processing. The time calibration module performs noise suppression and time calibration based on these anchor points, thereby eliminating the problem of timestamp deviation between different detection devices and ensuring the consistency and accuracy of multi-source data in the time dimension. The feature extraction module can extract key feature information of tongue images, pulse conditions, and thermal imaging from the calibrated data and convert it into feature vectors or parameter sets for analysis. The evolution index construction module further converts this feature information into a time series index sequence, capturing the dynamic change law of pediatric physiological data over time. The visualization module constructs visual coding rules for pediatric data based on these time series index sequences and performs intuitive visual display, enabling the clear observation of the evolution process of tongue image, pulse condition, and body surface temperature data. In summary, the present invention can effectively solve the problem of dynamic correlation analysis caused by the heterogeneity of multi-dimensional physical sign data in the clinical data processing of traditional Chinese medicine pediatrics.

[0013] Preferably, the collection of pediatric physiological time series data sets includes: Collecting multi-dimensional physiological information of children through multi-source traditional Chinese medicine pediatric detection devices to obtain a multi-modal traditional Chinese medicine pediatric physiological data set; Carrying out device identification code registration for multi-source traditional Chinese medicine pediatric detection devices to obtain a device identification code data table; using a preset network time protocol server to perform clock synchronization calibration on multi-source traditional Chinese medicine pediatric detection devices to obtain a device reference time synchronization record; Performing data source parsing on the multi-modal traditional Chinese medicine pediatric physiological data set according to the device identification code data table and the device reference time synchronization record to obtain a device data stream meta-information table; Extracting the timestamp field from the device data stream meta-information table and performing standardized conversion to obtain an original acquisition timestamp sequence; calculating the time deviation value between each detection device based on the original acquisition timestamp sequence to obtain a device time difference parameter table; Calculating the standard deviation of the time deviation of 5 consecutive measurements in the device time difference parameter table, and performing stability judgment according to the standard deviation of the time deviation and a preset stability threshold to obtain a time deviation stability judgment result; Quantifying the reliability of the device time difference parameter table according to the time deviation stability judgment result to obtain a time difference reliability evaluation result; performing weighted average correction on the device time difference parameter table according to the time difference reliability evaluation result to obtain a corrected device time difference parameter table; Performing dynamic time warping on the multi-modal traditional Chinese medicine pediatric physiological data set based on the corrected device time difference parameter table to obtain a pediatric physiological time series data set, where the pediatric physiological time series data set includes pulse waveform data, tongue coating microscopic image sequences, and body surface thermal imaging data.

[0014] In this embodiment, during a health check of a child in a pediatric health care clinic, a pulse condition detector, a tongue image detector, and a body surface thermal imager are used to collect pulse waveform data, a sequence of tongue coating microscopic images, and body surface thermal imaging data respectively. Among them, the pulse condition detector uses an intelligent pulse diagnosis instrument - PM3000. This device is attached to the radial artery of the child's wrist through a pressure sensor and continuously collects pulse waveform signals for 10 minutes at a sampling frequency of 1000 Hz. The signal amplitude range is 0 - 5V. The tongue image detector selects an intelligent tongue diagnosis instrument - TS2000, which is equipped with a high-definition camera and a standardized light source. When taking pictures of the child's tongue, a sequence of tongue coating microscopic images with a resolution of 2048×1536 pixels is obtained at a shooting distance of 30 cm. 2 frames are taken per second, and a total of 5 minutes of image data is collected. The body surface thermal imager uses a medical-grade infrared thermal imager - TH9000, whose infrared detector has a resolution of 640×480 pixels and a temperature detection range of 30℃ - 45℃. At a distance of 1 meter from the child's body surface, body surface thermal imaging data is collected at a frequency of 1 frame per second for a duration of 10 minutes. Through these three devices, a multi-modal traditional Chinese medicine pediatric physiological data set containing pulse waveform data, a sequence of tongue coating microscopic images, and body surface thermal imaging data is finally obtained and stored in the medical data server in the clinic. Use Microsoft Excel to create a device identification code data table, and assign unique device identification codes: PM-001, TS-002, and TH-003 to the pulse condition detector, the tongue image detector, and the body surface thermal imager respectively, and record the device information (including device model, purchase date, device serial number) in the table. At the same time, connect a dedicated time server configured with an NTP (Network Time Protocol) server to the local area network in the clinic. This server is synchronized with the time source of the National Time Service Center through the Internet, and the time synchronization accuracy reaches the millisecond level. Connect the three detection devices to the local area network through network cables, and configure NTP clients in their device management systems respectively to synchronize their clocks with the time server. After calibration, the device reference time synchronization record shows that the time deviation of the pulse condition detector is 1.2 milliseconds, the time deviation of the tongue image detector is 0.8 milliseconds, and the time deviation of the body surface thermal imager is 1.5 milliseconds, all within the acceptable synchronization accuracy range. Import the collected multi-modal traditional Chinese medicine pediatric physiological data set into the data processing workstation, and the medical data management software Health Data manager is running on this workstation. In the software, the data source of the data set is parsed based on the previously established device identification code data table and the device reference time synchronization record. In the Health Data manager software, associate the pulse data file with the device identification code PM-001, associate the tongue image sequence file with TS-002, and associate the body surface thermal imaging data file with TH-003. At the same time, input the time synchronization calibration parameters (such as time deviation values) of each device into the software.Based on this information, the software automatically generates a metadata table for the device data stream, which details the device identification code, data type, acquisition start time, acquisition duration, data format, and time synchronization calibration parameter information corresponding to each data file. In the Health Datamanager software, a timestamp field extraction operation is performed on each data file in the device data stream metadata table. For the pulse condition data file, the software reads its timestamp field and finds that its time format is the relative time since the device was powered on (unit: second), and the time accuracy is at the millisecond level. Using the built-in timestamp standardization conversion function module in the software, this relative timestamp is converted into an absolute time format that conforms to the ISO 8601 standard (e.g., 2025-02-20T14:30:25.123+08:00), which includes the time zone information (+08:00 represents the eighth time zone). Similarly, timestamp extraction and conversion operations are performed on the tongue image sequence file and the body surface thermal imaging data file respectively, and finally an original acquisition timestamp sequence file is obtained, which records the precise acquisition time of each data point or image frame. Using the Health Data manager software, the time deviation values between the detection devices are calculated based on the original acquisition timestamp sequence. Taking the pulse condition detector as the reference, the time deviations between the tongue image detector and the body surface thermal imager and the pulse condition detector are calculated respectively. The software extracts the timestamp data of each device from the original acquisition timestamp sequence every 1 minute, and calculates the time deviation sequence between the tongue image detector and the pulse condition detector. For example, in the first 5 measurements, the time deviations are 0.5 milliseconds, 0.6 milliseconds, 0.5 milliseconds, 0.4 milliseconds, and 0.5 milliseconds respectively; the time deviation sequences between the body surface thermal imager and the pulse condition detector are 1.0 milliseconds, 1.2 milliseconds, 1.1 milliseconds, 1.3 milliseconds, and 1.0 milliseconds respectively. The software calculates the standard deviation of the time deviations for 5 consecutive measurements of these two time deviation sequences. For the tongue image detector, the standard deviation is approximately 0.05 milliseconds; for the body surface thermal imager, the standard deviation is approximately 0.10 milliseconds. The preset stability threshold is set to 0.2 milliseconds. Therefore, the software determines that the time deviations between the tongue image detector and the body surface thermal imager and the pulse condition detector are both in a stable state. The generated time deviation stability judgment results show that the time deviation stabilities of both are stable, and the results are stored in the device time difference parameter table. Based on the time deviation stability judgment results, the reliability of the device time difference parameter table is quantified in the Health Data manager software. For the device time deviation data with stable stability judgment, the software assigns a higher weight value (e.g., the weight of the tongue image detector is 0.8, and the weight of the body surface thermal imager is 0.7), while for the data with poor stability (assuming there are other devices), the weight will be correspondingly reduced. The software performs weighted average correction on the time deviation data in the original device time difference parameter table according to these weight values.For example, for the tongue image detector, after weighted average correction, the time deviation correction value from the pulse detector is 0.52 milliseconds; for the body surface thermal imager, the corrected time deviation value is 1.06 milliseconds. Using the corrected device time difference parameter table, dynamic time warping is performed on the multi-modal pediatric TCM physiological data set in the Health Data manager software. The software first uses the time axis of the pulse waveform data collected by the pulse detector as a reference, and adjusts the time axes of the tongue image microscopic image sequence and the body surface thermal imaging data according to the corrected time difference parameters. For the tongue image sequence, the software fine-tunes the acquisition time of each image frame forward or backward according to a time deviation of 0.52 milliseconds, so that the time axis of the image sequence is aligned with the time axis of the pulse waveform data. Similarly, the time of each image frame of the body surface thermal imaging data is adjusted accordingly according to a deviation of 1.06 milliseconds. After dynamic time warping, a pediatric physiological time series data set with a unified time reference is obtained, which includes calibrated pulse waveform data, tongue coating microscopic image sequences, and body surface thermal imaging data.

[0015] In the present invention, by performing unified device identification coding registration and network time protocol calibration on multi-source pediatric TCM detection devices, the data acquisition process is standardized from the source to ensure the consistency of the time reference of each device. Through detailed data source analysis and timestamp standardization conversion of the multi-modal data set, the time deviation parameters between devices are accurately calculated, and the stability and reliability of the time deviation are further quantified, enabling the system to accurately identify and correct time deviation problems. By performing dynamic time warping on the original multi-modal data based on the corrected device time difference parameter table, a unified pediatric physiological time series data set is generated, effectively integrating the physiological information of pulse waveforms, tongue coating microscopic images, and body surface thermal imaging, providing a high-quality and time-consistent data basis for subsequent feature extraction and analysis, thus significantly improving the efficiency and credibility of the entire pediatric TCM data processing system.

[0016] Preferably, the identification of physiological event anchor points for the pediatric physiological time series data set includes: Performing real-time wave peak detection on the pulse waveform data to obtain a pulse wave peak position sequence; Calculating the time interval between adjacent wave peaks according to the pulse wave peak position sequence, and performing physiological validity screening on the pulse wave peak position sequence according to the time interval between adjacent wave peaks and the preset physiological cycle threshold range to obtain a standard pulse cycle marker point set; Performing frame-by-frame comparison on the tongue coating microscopic image sequence based on the frame difference method to obtain tongue image rapid change frame marker points; Calculating and marking the regional temperature gradient of the body surface thermal imaging data to obtain a temperature sudden change region marker map; Statistically determine the time points of the temperature abrupt change regions in consecutive frames of body surface thermal imaging data based on the temperature abrupt change region marking map, and generate a sequence of thermal imaging event trigger time points; Record the corresponding event trigger points in the pulse wave peak position sequence, the marked points of the rapidly changing frames of tongue images, and the thermal imaging event trigger time point sequence as the initial physiological event anchor point set; Construct a cross-modal event time correspondence table based on the pulse wave peak position sequence, the marked points of the rapidly changing frames of tongue images, and the thermal imaging event trigger time point sequence; Make the following judgments on each event anchor point in the initial physiological event anchor point set based on the cross-modal event time correspondence table: If trigger events exist in at least two modalities within the ±200 ms window, mark the corresponding event anchor point as a high-confidence anchor point; otherwise, eliminate the corresponding event anchor point, and record the high-confidence anchor points as the physiological event anchor point set.

[0017] In this embodiment, the Health Signal Analyzer software is used to perform real-time analysis on the pulse waveform data. The built-in peak detection algorithm of the software is based on the sliding window technique. The window length is set to 0.5 seconds and the step size is 0.1 seconds. When the pulse waveform data is input, the algorithm calculates the first derivative within each sliding window, and locates the positions where the derivative crosses zero and the waveform amplitude exceeds 2 standard deviations of the baseline as candidate peaks. Further, the valid peaks with peak-to-peak intervals between 0.5 - 1.5 seconds are selected, and finally the pulse peak position sequence is obtained. The sequence records the timestamps and amplitudes corresponding to each peak. For example, it is detected that the peaks appear at positions such as the 5th second, 10.2 seconds, 15.3 seconds, etc., and the amplitudes are 2.3 mV, 2.1 mV, 2.4 mV respectively. In the Health Signal Analyzer software, the physiological cycle analysis function module of the software is used to perform validity screening on the obtained pulse peak position sequence. The preset physiological cycle threshold range of the software is 0.4 - 1.6 seconds (corresponding to a heart rate range of 37.5 - 150 beats per minute, which is within the normal heart rate range for children). The software first calculates the time intervals between adjacent peaks according to the pulse peak position sequence. For example, the time intervals between the first three peaks are 5.2 seconds and 5.1 seconds respectively. Then, by comparing with the preset threshold, it is found that these intervals all exceed the threshold upper limit of 1.6 seconds, indicating an abnormality. Further inspection reveals that the child had limb movement interference during the detection process, resulting in deviations in the detection of some peaks. The parameters are adjusted to narrow the sliding window to 0.3 seconds, and peak detection and screening are performed again. Finally, a set of standard pulse cycle marker points with adjacent peak time intervals between 0.6 - 1.4 seconds is obtained. The timestamps of the marker points are at the 8th second, 13.2 seconds, 18.5 seconds, etc., and the corresponding heart rates are 43 - 95 beats per minute, which is within the normal heart rate fluctuation range of the child in a quiet state. The Tongue Image Analyzer software is used, and this software performs image sequence analysis based on the inter-frame difference method. The inter-frame difference threshold is set to 0.2 (the value after normalization) in the software, that is, when the gray difference between two adjacent frames of images exceeds this threshold, it is determined that the tongue image has changed. The software compares the tongue image sequence frame by frame, and it is detected that the inter-frame difference between the 12th frame and the 13th frame is 0.25, exceeding the set threshold, and the 13th frame is marked as a frame with a rapid change in the tongue image. Further inspection reveals that when the child slightly opens the mouth or wiggles the tongue, the tongue image will change. The software has marked a total of 15 frames with rapid changes in the tongue image, and the timestamps are distributed at positions such as the 5th second, 14 seconds, 23 seconds, etc. in the acquisition sequence. The Thermal Scan Processor software is used to analyze the surface thermal imaging data of the child. The software first preprocesses the thermal imaging data, including dead pixel correction and temperature calibration. In the temperature gradient calculation module, the spatial gradient threshold is set to 0.5 °C / cm and the time gradient threshold is set to 0.3 °C / second.The software calculates the temperature difference between each pixel of the body surface thermal imaging data and its 8 surrounding neighboring pixels. If the temperature difference exceeds 0.5 °C / cm, it is marked as a region of sudden temperature change. At the same time, the software compares the temperature change rate of the pixels at the same position in consecutive frames. If it exceeds 0.3 °C per second, it is also marked as a region of sudden temperature change. After calculation, the software generates a marked map of the regions of sudden temperature change, where the regions of sudden temperature change are identified by different colors. For example, a sudden temperature change appears in the periumbilical region of the abdomen of the child, indicating a change in gastrointestinal function; sudden temperature changes are also detected on the forehead and neck, which are related to the fever state. Based on the marked map of the regions of sudden temperature change, the ThermalScan Processor software further counts the time points of the regions of sudden temperature change in consecutive frames of the body surface thermal imaging data. The software sets the shortest continuous time of sudden temperature change to 3 frames (corresponding to 0.3 seconds) to avoid interference from instantaneous noise. When a region of sudden temperature change appears in consecutive frames, the time stamp of the starting frame is recorded. For example, during the period from 7 seconds to 7.3 seconds, a sudden temperature change continuously appears in the periumbilical region of the child's abdomen, and the software marks 7 seconds as the trigger time point of the thermal imaging event; from 15 seconds to 15.4 seconds, a similar situation occurs in the forehead region, and the trigger time point is also recorded. The finally generated sequence of trigger time points of thermal imaging events contains 8 time points, distributed at positions such as 7 seconds, 15 seconds, and 22 seconds in the acquisition sequence. The sequence of pulse wave peak positions, the marked points of the rapidly changing frames of tongue images, and the sequence of trigger time points of thermal imaging events are imported into the Health Data Integrator software. This software has a cross-modal data fusion function and can read the time stamp information of different data types. In the time axis alignment interface of the software, the time axes of the three modalities of data are visually compared. For example, the pulse wave peaks appear at 8 seconds and 13.2 seconds; the rapidly changing frames of tongue images are at 5 seconds and 14 seconds; the thermal imaging events are triggered at 7 seconds and 15 seconds. The software automatically associates the event trigger points at the corresponding time points in the three modalities of data to construct an initial set of physiological event anchor points. The set of anchor points is presented in the form of a list, and each anchor point contains the event type (pulse image, tongue image, thermal imaging), time stamp, and related characteristic values. For example, anchor point 1: pulse wave peak, time stamp 8 seconds, amplitude 2.3 mV; anchor point 2: tongue image change, time stamp 5 seconds, frame difference 0.25; anchor point 3: thermal imaging event, time stamp 7 seconds, temperature change rate 0.4 °C per second. In the Health Data Integrator software, a cross-modal event analysis module is used to construct an event time correspondence table. The software first extracts the time stamp information from the initial set of physiological event anchor points and arranges them in chronological order. For example, the anchor point time stamps are 5 seconds (tongue image), 7 seconds (thermal imaging), 8 seconds (pulse image), 13.2 seconds (pulse image), 14 seconds (tongue image), 15 seconds (thermal imaging), etc. The software creates a time correspondence entry for each event, recording the event type, time stamp, and modal source.Meanwhile, the software calculates the time intervals between adjacent events. For example, the interval between 5 seconds and 7 seconds is 2 seconds, and the interval between 7 seconds and 8 seconds is 1 second, etc. The generated cross-modal event time correspondence table is presented in tabular form, including columns for event ID, timestamp, event type, modal source, and adjacent event interval. The distribution and correlation of different modal events on the time axis can be visually observed through the table. Based on the cross-modal event time correspondence table, the Health Data Integrator software activates the high-confidence anchor screening function. The software sets the time window to ±200 milliseconds and judges each event anchor in the initial physiological event anchor set. For example, for the pulse wave peak anchor with a timestamp of 8 seconds, the software checks whether there are events of other modalities within the range of 200 milliseconds before and after it. It is found that there is an anchor for the rapidly changing frame of the tongue image at 8.1 seconds, which is outside the time window range (8 seconds + 200 milliseconds = 8.2 seconds), so it does not meet the conditions; but there is an anchor for the thermal imaging event at 7.8 seconds, which is within the time window. Therefore, the software marks the pulse wave peak anchor at 8 seconds as a high-confidence anchor. After screening all the initial anchors, the software finally determines 12 high-confidence anchors, where events in at least two modalities among the three modalities appear simultaneously within ±200 milliseconds, ensuring the reliability of event correlation. The high-confidence anchor set is exported as a.csv file, containing event ID, timestamp, and associated modal type information.

[0018] The present invention ensures the accuracy of pulse cycle marking by performing real-time wave peak detection on pulse waveform data and combining physiological cycle thresholds for effective screening. By analyzing the tongue coating microscopic image sequence through the inter-frame difference method, the rapidly changing frames of the tongue image are captured, and the temperature gradient calculation of the body surface thermal imaging data can accurately locate the regions and time points of sudden temperature changes. After integrating the pulse wave peak positions, tongue image change frames, and thermal imaging event trigger time points into the initial physiological event anchor set, further screening of high-confidence anchors is carried out based on the cross-modal event time correspondence table, effectively removing false event anchors caused by noise or equipment errors. This not only improves the accuracy and reliability of physiological event anchor recognition but also provides an accurate time reference for subsequent time calibration and feature extraction, enabling the system to more effectively capture and analyze key events in pediatric physiological data.

[0019] Preferably, the time calibration of the noise-suppressed physiological time series dataset includes: Segmenting the noise-suppressed physiological time series dataset with a fixed window length of 10 seconds and an adjacent window overlap rate of 20% to generate a segmented time series dataset; Calculating the slope and intercept parameters of each segment of data in the segmented time series dataset to generate a segmented linear parameter table; Calculate the mean square residual of each segment of data in the segmented time series dataset according to the piecewise linear parameter table, and determine and screen the piecewise linear parameter table according to the mean square residual and the preset fitting quality screening threshold to generate a valid piecewise linear parameter table; Calculate the difference of the slopes of adjacent segments in the valid piecewise linear parameter table to generate an adjacent segment slope difference sequence; Identify the slope mutation points according to the adjacent segment slope difference sequence to generate a drift mode switching point sequence; Perform pattern matching on the valid piecewise linear parameter table according to the preset typical drift modes and the drift mode switching point sequence to generate a drift mode classification table; Perform adaptive smoothing on the connection points of different mode junction types based on the drift mode classification table to generate a piecewise drift characteristic curve, where the adaptive smoothing is specifically: When the junction type is an increasing / decreasing mode junction, cubic Hermite interpolation is used for smoothing; When the junction type is an oscillating mode junction, mean smoothing filtering is used for smoothing; Perform cubic spline interpolation on the piecewise drift characteristic curve to generate a signal time drift characteristic curve; construct a time mapping rule according to the signal time drift characteristic curve; Perform time coordinate transformation on the noise-suppressed physiological time series dataset based on the time mapping rule to obtain a time-corrected physiological dataset, where the time-corrected physiological dataset includes calibrated pulse waveform data, calibrated tongue coating microscopic image sequences, and calibrated body surface thermal imaging data.

[0020] In this embodiment, the signal processing software Signal Pro X is used to perform time calibration preprocessing on the collected physiological time series dataset with noise suppression. In the software, 10 seconds is set as the fixed window length, and the overlapping rate between adjacent windows is 20% (that is, there is an overlapping part of 2 seconds between adjacent windows). Through the data segmentation function module of the software, the original pulse waveform dataset is segmented into multiple small segments, each segment containing 1000 sampling points (10 seconds × 100 Hz), and 200 sampling points are shared between adjacent segments. The finally generated segmented time series dataset contains 360 segments, and these segments are stored in a dedicated folder, and each segment is saved in the ".seg" format. Analyze the segmented time series dataset in the Signal Pro X software. The linear fitting parameter extraction function module of the software is used to calculate the slope and intercept parameters of each segment of data. This module uses the least squares fitting algorithm to perform linear fitting on the pulse waveform data within each 10-second window. For example, the slope of the fitting line for the first segment is 0.05 mV / s, and the intercept is 1.2 mV; the slope of the second segment is 0.03 mV / s, and the intercept is 1.5 mV. These slope and intercept parameters are automatically organized into a table file, that is, the segmented linear parameter table, which contains the start time, end time, slope, and intercept values of each segment. After obtaining the segmented linear parameter table, use Microsoft Excel to determine the fitting qualification of the data. First, according to the slope and intercept values in the segmented linear parameter table, use the formula function of Excel to calculate the mean square residual of each segment of data. The calculation formula for the mean square residual is: , where is the actual data point, is the predicted value corresponding to the fitted line, and n is the number of data points. In Excel, the fitting quality screening threshold is set to a mean square residual less than 0.1 mV². Through the screening function, segments with a mean square residual exceeding the threshold are excluded, and finally an effective piecewise linear parameter table containing 280 segments is obtained. Based on the effective piecewise linear parameter table, the Signal Pro X software is used again to call its data sequence analysis function module. This module can read the slope sequence in the effective piecewise linear parameter table and calculate the difference between adjacent segment slopes. For example, if the slope of the first effective segment is 0.05 mV / s and the second is 0.03 mV / s, the difference is -0.02 mV / s²; the difference in slope between the second and third segments is +0.01 mV / s², and so on. The software automatically generates a file of the difference sequence of adjacent segment slopes, which records the slope change rate between each adjacent segment pair. In the Signal ProX software, the mutation point detection function module is used to analyze the difference sequence of adjacent segment slopes. This module uses a method based on a statistical threshold to identify slope mutation points. According to experience, the absolute value threshold of the slope change rate is set to 0.02 mV / s². The software traverses the difference sequence, and when a difference value exceeding this threshold is detected, the corresponding position is marked as a drift mode switching point. For example, at the 50th segment pair of the sequence, the difference value reaches -0.03 mV / s², exceeding the threshold and being identified as a drift mode switching point. After processing, the software generates a file of the drift mode switching point sequence, which records the positions of all switching points in the original data and the corresponding slope change rates. The drift mode switching point sequence is imported into the pattern recognition function module of the Signal Pro X software. This module has several typical drift mode templates built-in, including an increasing mode, a decreasing mode, and an oscillating mode. The software compares the segment data in the effective piecewise linear parameter table with the preset typical drift modes through a pattern matching algorithm. For example, when the slopes of three consecutive segments are 0.05 mV / s, 0.04 mV / s, and 0.03 mV / s in sequence, it is matched to the decreasing mode; if the segment slopes alternate between positive and negative, such as 0.02 mV / s, -0.01 mV / s, 0.03 mV / s, it is matched to the oscillating mode. The software automatically generates a drift mode classification table, which shows the drift mode type to which each segment belongs in tabular form and marks different modes with different colors. Based on the drift mode classification table, the adaptive smoothing function module is called in the Signal Pro X software. For the segment connection where the junction type is the increasing / decreasing mode junction, the software uses the cubic Hermite interpolation method for smoothing. For example, at the junction of the increasing mode and the decreasing mode, the software calculates the interpolation points for smooth transition according to the slopes and data points of adjacent segments. For the connection where the junction type is the oscillating mode junction, the mean smoothing filter method is used, and the average value of adjacent data points is calculated to reduce the fluctuations caused by oscillations.After adaptive smoothing, the software generates a segmented drift characteristic curve. Using the advanced interpolation function module of Signal Pro X software, cubic spline interpolation is performed on the segmented drift characteristic curve. The software constructs a smooth and continuous signal time drift characteristic curve based on the key points in the segmented drift characteristic curve, such as the drift mode switching points and the data points after smoothing. For example, in the increasing mode segment of the curve, cubic spline interpolation can accurately depict the slow rising trend of the signal; in the oscillation mode segment, the interpolated curve can better fit the fluctuation characteristics of the original data. Based on the generated signal time drift characteristic curve, the software automatically constructs a time mapping rule. This rule defines the conversion relationship between the original time coordinate and the corrected time coordinate. For example, the signal value at the original time point t will be mapped to the corrected time point t’ = t + Δt (Δt is the time offset calculated according to the drift characteristic curve). The time mapping rule is stored in the calibration parameter file of the software in the form of a function or a lookup table. Applying the constructed time mapping rule in Signal Pro X software, a time coordinate transformation is performed on the noise-suppressed physiological time series dataset. The software reads the original pulse waveform data, tongue coating microscopic image sequence, and body surface thermal imaging data, and adjusts the timestamp of each data point according to the conversion relationship in the time mapping rule. For example, a sampling point with an original time of 100 seconds in the pulse waveform data is corrected to 100.5 seconds after being transformed by the time mapping rule. After processing, a time-corrected physiological dataset is obtained, which includes calibrated pulse waveform data, calibrated tongue coating microscopic image sequence, and calibrated body surface thermal imaging data.

[0021] In the present invention, by performing segmented processing on the data and calculating the slope and intercept of each segment, a segmented linear parameter table is generated, providing a basis for subsequent analysis. Through the determination of the mean square residual and the fitting quality screening threshold, the reliability of the segmented linear parameters is ensured. By calculating the difference of adjacent segmented slopes and identifying the slope mutation points, the key points of data drift can be accurately located. Based on the preset typical drift modes for pattern matching and classification, the recognition ability of data drift characteristics is further improved. At the junction of different drift modes, adaptive smoothing techniques are adopted. For example, cubic Hermite interpolation is used at the junction of the increasing / decreasing mode, and mean smoothing filtering is used at the junction of the oscillation mode, effectively eliminating the discontinuity and noise interference in the data. Through the signal time drift characteristic curve generated by cubic spline interpolation, an accurate time mapping rule is constructed, realizing the time coordinate transformation of the noise-suppressed physiological time series dataset. This not only improves the time consistency of the data but also ensures the accurate alignment of the calibrated physiological dataset in the time dimension.

[0022] Preferably, generating the tongue image feature vector set based on the tongue image set includes: Performing histogram equalization on the tongue image set to obtain an enhanced tongue image set; Perform semantic segmentation on the enhanced tongue image set to obtain a tongue body region mask image; extract regions from the enhanced tongue image set according to the tongue body region mask image to obtain a tongue body ROI image set; Perform color space conversion and decomposition on the tongue body ROI image set to obtain a tongue image feature channel group; Perform wavelet transform on the tongue image feature channel group to obtain a tongue image texture wavelet coefficient table; extract statistical features from the tongue image texture wavelet coefficient table to obtain a tongue image texture feature table; Calculate local binary pattern features based on the tongue body ROI image set to obtain a tongue image LBP feature map; analyze the color distribution features according to the tongue body ROI image set to obtain a tongue color feature vector; Fuse the tongue image texture feature table, the tongue image LBP feature map and the tongue color feature vector to obtain a tongue image feature vector set.

[0023] In this embodiment, the tongue image set is imported into the image processing software Image J. In Image J, the histogram equalization function module is selected, and the parameters of histogram equalization are set: the gray level range is set to 0 - 255, and the default global equalization method is adopted. The software performs histogram equalization processing on the R, G, and B channels of each tongue image respectively. After processing, the contrast of the tongue image is significantly improved, and the details of the tongue body, tongue coating, and tongue color are clearer. For example, in an original image, the color of the tongue coating is darker and the boundary with the tongue body color is blurred. After equalization, the texture and color levels of the tongue coating are more distinct. The enhanced tongue image set is imported into the semantic segmentation software SegLab. This software is based on a deep learning semantic segmentation algorithm and pre-trains a model suitable for tongue image analysis. In the software, the tongue body segmentation model is selected, and the confidence threshold is set to 0.8. The software performs semantic segmentation processing on each enhanced tongue image and outputs a binary mask image of the tongue body region. In the mask image, the tongue body region is displayed in white (pixel value is 255), and the background region is displayed in black (pixel value is 0). For example, in an image, the tongue body region is accurately covered by the white mask with clear edges. Using the region extraction function module of Image J software, the tongue body mask image and the enhanced tongue image are subjected to a pixel-level AND operation to extract the region of interest (ROI) of the tongue body. The extracted tongue body ROI image set is stored in a separate folder, and each image only contains the tongue body part, removing the background and other interference information. For example, the extracted image clearly shows the shape, color, and texture of the tongue body, with the resolution being the same as that of the original image, but only retaining the details of the tongue body part. Using the color space conversion plug-in of Image J software, the tongue body ROI image set is converted from the RGB color space to the LAB color space. The converted color space is selected as LAB, and the option to decompose the color channels is checked. The software performs conversion and decomposition operations on each tongue body ROI image to generate three single-channel images: L (luminance), A (green - red channel), and B (blue - yellow channel). For example, the L-channel image after conversion of a tongue body ROI image shows the light and dark changes of the tongue body, the A-channel highlights the difference in the red and green components of the tongue body, and the B-channel reflects the difference in the yellow and blue components. The decomposed tongue image feature channel group is stored in the form of a folder. Each folder corresponds to a color channel and contains the grayscale images of the corresponding channels of all tongue body ROI images. The Wavelet Tool box plug-in is installed in Image J software. The Daubechies wavelet basis function (db4) is selected, and the decomposition level is set to 3 layers. Wavelet transform is performed on each channel image in the tongue image feature channel group respectively. Taking the A-channel image as an example, after wavelet transform, the approximation coefficient and detail coefficient matrices are obtained and stored as the tongue image texture wavelet coefficient table. The table records the wavelet coefficient values at different scales. For example, the value of the approximation coefficient at the third layer is 0.75, and the value of the detail coefficient at the first layer is 0.32.Use the statistical analysis function module of Image J to extract statistical features from the wavelet coefficient table. Calculate the mean, standard deviation, skewness, and kurtosis of the wavelet coefficients of each channel image. For example, the mean of the wavelet coefficients of channel A is 0.52, and the standard deviation is 0.18. These statistical features form part of the tongue image texture feature table. Finally, the tongue image texture feature table integrates the statistical features of the L, A, and B channels. In the Image J software, use the Local Binary Patterns plugin to calculate the local binary pattern features of the tongue ROI image set. Set the neighborhood radius of the LBP operator to 3 pixels and the number of sampling points to 8, and use the default rotation-invariant and gray-scale invariant modes. The software calculates the LBP features for each tongue ROI image and generates a tongue image LBP feature map. The map is displayed in the form of a grayscale image, and each pixel value represents the LBP pattern coding value at that position. For example, in an LBP feature map, the tongue coating area shows a relatively high texture pattern coding value, reflecting the microscopic structural features of the tongue coating. At the same time, use the color histogram function module of ImageJ to analyze the color distribution features of the tongue ROI image set. The software calculates the histograms of the R, G, and B channels of each image, counts the number of pixels at each gray level, and performs normalization. For example, the histogram of the red channel of the tongue image shows a relatively high pixel proportion in the gray level range of 180 - 255, corresponding to the color characteristics of the tongue body. According to the histogram statistical results, calculate the color moments (mean, variance, skewness) as components of the tongue color feature vector. For example, the color mean of the red channel is 0.72, and the variance is 0.08. These values form part of the elements of the tongue color feature vector. Use the feature fusion software Feature Fusion Tool. In the software, use the tongue image texture feature table (including the mean, standard deviation, skewness, and kurtosis statistical features of the L, A, and B channels, a total of 16-dimensional features), the tongue image LBP feature map (the LBP histogram features obtained through statistics, a total of 256-dimensional features), and the tongue color feature vector (including the color moments of the R, G, and B channels, a total of 9-dimensional features) as inputs. The software provides a variety of feature fusion methods. Select the linear fusion method and set the weights of each feature group: the weight of the tongue image texture feature table is 0.4, the weight of the tongue image LBP feature map is 0.4, and the weight of the tongue color feature vector is 0.2. The software adds the elements of each feature vector according to the set weights to generate a set of tongue image feature vectors. For example, the dimension of the fused feature vector is 16 + 256 + 9 = 281 dimensions, and each feature vector corresponds to the comprehensive feature description of a tongue ROI image. The set of fused tongue image feature vectors is stored in the output file of the software in matrix form, where each row represents the feature vector of a tongue image, and the columns correspond to different feature dimensions.

[0024] The present invention enhances the contrast of tongue image through histogram equalization, making the image features more obvious. Through the application of semantic segmentation, the tongue region can be accurately separated from the image background, providing accurate regional positioning for subsequent feature extraction. Through color space transformation and decomposition operations, the image information is further decomposed into multiple feature channels, facilitating feature analysis from different dimensions. By using wavelet transform and local binary pattern feature calculation, key features are extracted from two aspects of texture and color distribution respectively, ensuring that the feature vector can comprehensively reflect the visual information of the tongue image. By fusing tongue texture features, LBP features and tongue color feature vectors, the generated tongue image feature vector set not only has rich dimensions, but also has stronger ability to capture the detailed features of the tongue image.

[0025] Preferably, constructing a pulse characteristic parameter set based on the pulse signal data set includes: Performing pulse period detection and segmentation on the pulse signal data set to obtain a single-cycle pulse segment set; Performing continuous wavelet transform on the single-cycle pulse segment set to obtain a pulse time-frequency spectrum atlas; Extracting frequency band energy features based on the pulse time-frequency spectrum atlas to obtain a pulse frequency domain feature set; Calculating time-domain feature parameters for the noise-reduced pulse waveform data to obtain a pulse time-domain feature table; Constructing a pulse characteristic parameter set according to the pulse frequency domain feature set and the pulse time-domain feature table.

[0026] In this embodiment, the Pulse Analyzer Pro software is used to analyze the collected pulse signal dataset. The built-in pulse period detection module in the software adopts the adaptive threshold method to identify the starting point and ending point of the pulse wave. The initial threshold is set to 20% of the signal amplitude, the minimum pulse period is 0.5 seconds, and the maximum pulse period is 2 seconds. The software automatically detects each pulse period and segments the continuous pulse signals into single-period pulse segments. For example, in the dataset, the first pulse period is detected to start from the 100th sampling point and end at the 300th sampling point; the second pulse period starts from the 301st sampling point and ends at the 500th sampling point, and so on. Finally, a set of 200 single-period pulse segments is obtained, and each segment is stored as an independent file. The single-period pulse segment set is imported into the Wavelet Tool box software. This software is a signal processing tool widely used in the field of biomedical signal analysis. The Morlet wavelet is selected as the mother wavelet function, and the scale range is set to 1 - 50 with a step size of 0.5. The software performs continuous wavelet transform on each single-period pulse segment to generate the pulse time-frequency spectrogram. For example, after transforming the first single-period pulse segment, the obtained time-frequency spectrogram shows the energy distribution of the pulse signal at different time points and frequencies. The time-frequency spectrogram is displayed in the form of a grayscale image, where the white area represents high energy and the black area represents low energy. Each file corresponds to the time-frequency spectrogram of a single-period pulse segment. In the Wavelet Toolbox software, the frequency band energy calculation function module is used to analyze the pulse time-frequency spectrogram set. The software divides the frequency range into four frequency bands: 0 - 2Hz, 2 - 5Hz, 5 - 10Hz, 10 - 20Hz. For each pulse time-frequency spectrogram, the software integrates and calculates the energy value within each frequency band. For example, in the first pulse time-frequency spectrogram, the energy of the 0 - 2Hz frequency band is 12.5, the energy of the 2 - 5Hz frequency band is 8.3, the energy of the 5 - 10Hz frequency band is 4.2, and the energy of the 10 - 20Hz frequency band is 1.8. These energy values are organized into a pulse frequency domain feature set, where each row represents the frequency band energy feature of a single-period pulse segment, and the columns correspond to different frequency bands. The denoised pulse waveform data is imported into the Signal Features software. Using the time domain feature extraction module of the software, multiple time domain feature parameters of the pulse waveform are calculated. The specific parameters include: peak value (maximum amplitude), valley value (minimum amplitude), rise time (time from the baseline to the peak), fall time (time from the peak back to the baseline), pulse interval (time between adjacent peaks), amplitude difference (difference between the peak and the valley), root mean square value (RMS), and waveform factor (ratio of the root mean square value to the average value).For example, in a section of denoised pulse waveform data, the calculated peak value is 2.3 mV, the trough value is 0.8 mV, the rise time is 0.12 seconds, the fall time is 0.25 seconds, the pulse interval is 0.85 seconds, the amplitude difference is 1.5 mV, the root mean square value is 1.2 mV, and the waveform factor is 1.3. These parameters are organized into a pulse time-domain feature table. Finally, using the Feature Integration Studio software, the pulse frequency-domain feature set and the pulse time-domain feature table are integrated, and the feature splicing method is selected to connect the frequency-domain features and time-domain features in sequence into a comprehensive feature vector. For example, the frequency-domain features of a single-cycle pulse segment are [12.5, 8.3, 4.2, 1.8] (corresponding to the energy of four frequency bands), and the time-domain features are [2.3, 0.8, 0.12, 0.25, 0.85, 1.5, 1.2, 1.3] (corresponding to the above eight time-domain parameters). After splicing, the comprehensive feature vector [12.5, 8.3, 4.2, 1.8, 2.3, 0.8, 0.12, 0.25, 0.85, 1.5, 1.2, 1.3] is obtained. The comprehensive feature vectors of all single-cycle pulse segments constitute the pulse characteristic parameter set, where each feature vector corresponds to the characteristic description of a single-cycle pulse segment.

[0027] Through pulse cycle detection and segmentation, the present invention can accurately extract single-cycle pulse segments, providing a basic unit for subsequent feature extraction. By continuous wavelet transform, the pulse signal is converted from the time domain to the time-frequency domain, and the generated pulse time-frequency spectrum atlas can intuitively display the time-frequency characteristics of the pulse signal, which helps to capture the subtle changes of the pulse. By extracting the frequency band energy characteristics, the energy distribution of the pulse in different frequency bands is further quantified, enriching the frequency-domain feature dimension of the pulse. By calculating the time-domain feature parameters, the core features of the pulse are supplemented from the time dimension, such as the amplitude and interval indicators of the pulse. The pulse characteristic parameter set constructed by integrating the frequency-domain features and time-domain features not only has rich dimensions and comprehensive information, but also can more accurately reflect the comprehensive features of the pulse, providing high-quality feature data for subsequent time series analysis and visualization, and effectively improving the recognition accuracy and analysis ability of the system for pulse features.

[0028] Preferably, generating the body area temperature feature table based on the thermal imaging image sequence includes: Performing temperature correction and pseudo-color mapping on the thermal imaging image sequence to obtain a standard body surface thermal imaging atlas; Performing threshold segmentation based on the standard body surface thermal imaging atlas to obtain a body surface area temperature mask map; Performing region division and marking on the body surface area temperature mask map to obtain a body part temperature region map; Calculating the temperature statistical features of each region based on the body part temperature region map to obtain the body area temperature feature table.

[0029] In this embodiment, the Thermal Image Processor software is used to preprocess the acquired sequence of body surface thermal imaging images. First, the software corrects the temperature of the original thermal imaging images. The temperature value of each pixel is adjusted through the built-in temperature calibration curve (provided by the device manufacturer, covering the range of 32°C to 42°C with an accuracy of ±0.1°C). The corrected temperature range is normalized to 34°C to 38°C. The software applies the pseudo-color mapping function to map the temperature values to the visible light color space, and selects the rainbow color mapping scheme, where 34°C corresponds to blue, 36°C corresponds to green, and 38°C corresponds to red. The processed standard body surface thermal imaging atlas is saved in the hospital's medical imaging archiving system in the ".png" format. The standard body surface thermal imaging atlas is imported into the image analysis software Image J. In the software, the threshold segmentation function is used to binarize the thermal imaging images. The temperature threshold range is set to 35°C to 37°C, and the pixels within this range are marked as the body surface area (white, pixel value 255), and the remaining background area is marked as black (pixel value 0). The processed body surface area temperature mask map is stored in a separate folder. The region analysis plugin of Image J is used to process the body surface area temperature mask map. The software automatically identifies and marks different body part regions. The specific operations are as follows: First, based on human anatomy knowledge, the initial marking points of the main body parts such as the head, neck, trunk, and limbs are manually marked on the first mask map. Based on these marking points, the software automatically divides the complete regions of each body part using the region growing algorithm (setting the similarity threshold to 0.85). For example, on a mask map, the head region is marked as label 1, the neck as label 2, the trunk as label 3, and the limbs are divided into labels 4 (left upper limb), 5 (right upper limb), 6 (left lower limb), and 7 (right lower limb). These images visually show the division results of different body parts. The body part temperature region map is re-imported into the Thermal Image Processor software, and temperature statistical analysis is performed in combination with the original thermal imaging data. The software automatically identifies each marked region and calculates its temperature statistical features. The specific statistical features include: average temperature, temperature standard deviation, temperature range (highest temperature - lowest temperature), temperature median, and temperature mode. For example, the analysis results of the trunk region (label 3) show an average temperature of 36.2°C, a standard deviation of 0.3°C, a range of 0.8°C, a median of 36.1°C, and a mode of 36.0°C. The temperature statistical features of all regions are organized into a table, where each worksheet corresponds to a body part, and the columns include timestamp, average temperature, standard deviation, range, median, and mode.

[0030] The present invention ensures accurate temperature data and better visual effects of thermal imaging images through temperature correction and pseudo-color mapping. The application of threshold segmentation enables the clear division of the temperature distribution in the body surface area, and the generated temperature mask map of the body surface area can accurately identify different temperature regions. Further regional division markings subdivide the body surface into specific body part temperature regions, making the temperature analysis more targeted and clinically significant. By calculating the temperature statistical features of each region, the generated body region temperature feature table not only provides comprehensive temperature data but also reflects the distribution law and change trend of the body surface temperature, providing high-quality feature data for subsequent body surface temperature rhythm analysis.

[0031] Preferably, constructing a tongue image time series index sequence based on the tongue image feature vector set includes: Performing color correction on the tongue image feature vector set to obtain a color-corrected tongue image feature vector set; Extracting the tongue body-tongue coating boundary features based on the color-corrected tongue image feature vector set to obtain a tongue body-tongue coating boundary contour sequence; Performing tongue image contour evolution simulation based on the tongue body-tongue coating boundary contour sequence to obtain tongue image contour evolution trajectory data; Calculating the tongue body stretching and contracting dynamic parameters based on the tongue image contour evolution trajectory data to obtain a tongue body movement index set; Performing wavelet packet decomposition on the color-corrected tongue image feature vector set to obtain a tongue image texture wavelet energy time series spectrum; Performing anisotropy analysis of the tongue coating roughness on the tongue image texture wavelet energy time series spectrum to obtain a tongue coating roughness-direction correlation index table; Calculating the regional humidity index of the tongue body-tongue coating boundary contour sequence to obtain a tongue body humidity index time series sequence; Generating a tongue image time series index sequence based on the tongue body movement index set, the tongue coating roughness-direction correlation index table, and the tongue body humidity index time series sequence.

[0032] In this embodiment, Adobe Photoshop software is used to perform color correction on the tongue image feature vector set. First, the tongue_roi_001.jpg image corresponding to the tongue image feature vector set is imported into Photoshop. By observing the histogram of the image, it is found that the color distribution is yellowish, which is caused by the lighting conditions during shooting. Select Image > Adjustments > Levels, and set the shadows, midtones, and highlights of the input levels to 0, 1.0, and 255 respectively to enhance the image contrast. Select Image > Adjustments > Color Balance, and in the midtones option, adjust the sliders for red, green, and blue to +15, -5, and +10 respectively to correct the problem of overall yellowish tone. Finally, select Image > Adjustments > Saturation, and increase the saturation to +10 to make the colors more vivid. The processed image is saved as tongue_corrected_001.jpg, and the corresponding feature vectors with updated color values are saved as tongue_vector_corrected_001.csv. Through the above operations, a tongue image color-corrected feature vector set is obtained. The color-corrected tongue image is imported into ImageJ software. In the software, select the edge detection plugin and use the Sobel operator for edge enhancement processing. Set the threshold to 0.2 (the normalized gradient magnitude), mark the edge points above the threshold as white (pixel value 255), and those below the threshold as black (pixel value 0). The preliminary boundary contour between the tongue body and the tongue coating is obtained through edge detection. Use the Analyze Particles function, set the size to 50 - 500 pixels and the circularity to 0.5 - 1.0, and perform morphological analysis on the edge to remove small noise points and pseudo-contours, and finally obtain a clear tongue body - tongue coating boundary contour sequence. In the tongue image contour evolution simulation step, MATLAB software is used. The tongue body - tongue coating boundary contour sequence is imported into MATLAB, and the active contour function in the Image Processing Toolbox is used for simulation. Set the number of evolution iterations to 100 times, and save the intermediate results every 10 iterations. For example, at the 10th iteration, slight changes begin to appear in the tongue image contour, and the contour lines gradually become smoother; by the 50th iteration, the contour changes tend to stabilize, showing the evolution trend of the tongue image contour. By observing the contour changes at different iteration times, tongue image contour evolution trajectory data is generated, including the contour coordinate information at each iteration. This data records the dynamic changes of the tongue image contour during the simulation process. Using the previously generated tongue image contour evolution trajectory data, a script is written in MATLAB to calculate the dynamic parameters of tongue body stretching and contraction. The script first extracts the contour coordinates from the trajectory data and calculates the area and perimeter of the tongue body at each time point. For example, at time point t1, the tongue body area is A1 and the perimeter is P1; at time point t2, the tongue body area is A2 and the perimeter is P2.By calculating the area change rate (such as (A2 - A1) / A1) and perimeter change rate (such as (P2 - P1) / P1) at adjacent time points, the telescopic dynamic parameters of the tongue body are obtained. These parameters are organized into a table, which contains columns such as timestamp, area, perimeter, area change rate, and perimeter change rate. Use the wavelet toolbox of MATLAB to perform wavelet packet decomposition on the tongue image color correction feature vector set. Select the db4 wavelet basis function and set the decomposition level to 3 layers. Decompose the RGB values of each pixel point in the tongue image feature vector to obtain wavelet packet coefficients. For example, for the RGB values of a pixel point (R = 120, G = 150, B = 80), wavelet packet coefficients at different frequencies and scales are obtained after decomposition. These coefficients reflect the contributions of the pixel point at different texture details. Organize the wavelet packet coefficients of all pixel points into a table, which contains pixel positions, RGB values, and corresponding wavelet packet coefficients. Import the tongue image texture wavelet energy time series spectrum into Image J software. Use the directional analysis plug-in of the software to perform anisotropic analysis on the spectrum diagram. Set the analysis directions to 0°, 45°, 90°, and 135°, and calculate the energy distribution in each direction. For example, in the tongue image texture wavelet energy time series spectrum, the total energy in the 0° direction is 1200, 800 in the 45° direction, 1000 in the 90° direction, and 900 in the 135° direction. According to these energy values, calculate the energy proportion in each direction to obtain the tongue coating roughness - direction correlation index table, which is saved in the form of an Excel file and contains information such as analysis direction, total energy, and energy proportion. Import the tongue body - tongue coating demarcation contour sequence into Image J software. Use the region analysis function of the software to calculate the humidity index for the tongue body region. According to the correlation model between tongue image color and humidity (this model is obtained based on a large number of clinical data statistics and stored in the software's database), the software automatically analyzes the RGB values of the tongue body color and converts them into a humidity index. For example, the average RGB values of the tongue body region in a tongue image are R = 180, G = 160, B = 130, and the humidity index calculated according to the model is 0.65 (the humidity index ranges from 0 to 1, and the higher the value, the greater the humidity). The calculation results of the humidity index are saved as an Excel file, which contains the timestamp of each tongue image and the corresponding humidity index value. Use SPSS Statistics software to integrate the tongue body movement index set, the tongue coating roughness - direction correlation index table, and the tongue body humidity index time series. First, align the three data sets according to the timestamp to ensure the consistency of data at each time point. Select the Data > Merge Files function to merge the three data sets into a comprehensive data set. In the merged data set, each time point contains tongue body movement parameters (area change rate, perimeter change rate), tongue coating roughness indexes (energy proportion in each direction), and tongue body humidity index. Use the Analyze > Descriptive Statistics function of SPSS to calculate statistics such as the mean and standard deviation of each index.The finally generated sequence of tongue image time series indicators contains comprehensive indicator data for all time points. This sequence comprehensively reflects the dynamic change characteristics of tongue images in the time series.

[0033] The present invention ensures the accuracy of the tongue image color information through color correction, providing a reliable basis for subsequent feature analysis. Through the extraction of the boundary characteristics between the tongue body and the tongue coating and the simulation of the evolution of the tongue image contour, the subtle changes in the tongue image morphology can be accurately captured, and the generated tongue image contour evolution trajectory data reflects the dynamic evolution process of the tongue body morphology. By calculating the dynamic parameters of the tongue body stretching and retracting, the movement characteristics of the tongue body are further quantified, providing an important indicator for analyzing the physiological state of the tongue body. Through wavelet packet decomposition and the anisotropy analysis of the tongue coating roughness, the texture characteristics of the tongue image are deeply explored, providing detailed information on the surface structure of the tongue coating. By calculating the regional humidity index, the change in the humidity level of the tongue image is reflected from another dimension. The finally generated sequence of tongue image time series indicators synthesizes the dynamic characteristics of the tongue body movement, the tongue coating texture, and the tongue body humidity, comprehensively reflecting the evolution law of the tongue image over time, and providing more detailed, quantitative, and dynamic feature data for the tongue diagnosis analysis in traditional Chinese pediatric medicine.

[0034] Preferably, generating the sequence of pulse image time series indicators based on the pulse characteristic parameter set includes: Performing zero-drift correction on the pulse characteristic parameter set to obtain a corrected pulse parameter time series sequence; Identifying the morphological characteristics of the main pulse wave based on the corrected pulse parameter time series sequence to generate a time series feature sequence of the main pulse wave morphology; Performing time-varying morphology clustering on the time series feature sequence of the main pulse wave morphology to generate a time-varying morphology clustering label set of the pulse wave; Constructing pulse image time-frequency combined feature tensor data based on the time-varying morphology clustering label set of the pulse wave; Calculating the pulse dynamic indicators for the pulse image time-frequency combined feature tensor data to generate a time series table of pulse dynamic indicators; Performing multi-lead signal synchronization according to the time series table of pulse dynamic indicators to generate a multi-lead pulse phase difference feature map; Calculating vascular elasticity parameters based on the multi-lead pulse phase difference feature map to generate a sequence of vascular elasticity indicators; Performing spatio-temporal evolution modeling on the sequence of vascular elasticity indicators to generate a sequence of pulse image time series indicators.

[0035] In this embodiment, the MATLAB software is used to perform zero-drift correction on the set of pulse characteristic parameters. Load the file of the set of pulse characteristic parameters, which contains multiple characteristic parameters of the pulse signal, such as pulse period, amplitude, and rise time. In MATLAB, call the detrend function in the Signal Processing Toolbox and select the constant option to remove the DC offset in the signal. For example, for a segment of the pulse signal, the original amplitude range is from 0.5V to 2.5V. After being processed by the detrend function, the zero-drift of the signal is corrected, and the new amplitude range becomes from -0.5V to 1.5V, and the baseline of the signal stabilizes near 0V. The corrected time series of the pulse parameters is saved as a new.mat file. Import the corrected time series of the pulse parameters into the Python environment and use the Biosppy library to identify the morphological features of the main wave of the pulse. Load the new.mat file and use the pulse_processing function in Biosppy. In the function, set the waveform type as radial (radial artery pulse), and adopt the default peak detection algorithm. For example, for a segment of the corrected pulse signal, the program identifies that the peak position of the main wave is at the 100th sampling point and the trough position is at the 150th sampling point. The feature extraction results include parameters such as the rising slope, falling slope, peak height, and trough depth of the main wave. These parameters are organized into a time series feature sequence of the morphological features of the main wave of the pulse and saved as a CSV file main_wave_features.csv. Use the scikit-learn library in Python to perform time-varying morphological clustering on the time series feature sequence of the morphological features of the main wave of the pulse. Load the main_wave_features.csv file, select the KMeans clustering algorithm, set the number of clusters to 3, and the initialization method to k-means++. For example, for the three features of the rising slope, falling slope, and peak height extracted, the program assigns the morphological feature vector of the main wave in each time window to the nearest cluster center. After multiple iterations, the program generates a time-varying morphological clustering label set for the pulse wave, and each label corresponds to a type of main wave morphology (such as label 0 represents the spike type, label 1 represents the gentle type, and label 2 represents the deep trough type). The clustering label sequence is saved as a CSV file pulse_cluster_labels.csv. In MATLAB, use the Signal Processing Toolbox to construct the time-frequency joint feature tensor data of the pulse. Load the original pulse signal file raw_pulse_signal.mat and the clustering label file pulse_cluster_labels.csv. Use the spectrogram function to perform time-frequency analysis on the pulse signal, set the window length to 1 second (corresponding to 1000 sampling points with a sampling frequency of 1000Hz), and the overlap length to 0.5 second.The calculated time-frequency map data is combined with the clustering labels to form three-dimensional tensor data, with dimensions of time, frequency, and clustering features respectively. For example, at time point t = 5 seconds and frequency f = 5 Hz, the tensor data contains the energy value at this frequency and the corresponding clustering label. The tensor data is saved as a.mat file named tensor_data.mat. Load the tensor_data.mat file and use the Tensor Toolbox in MATLAB to calculate the pulse dynamic indicators. Select the Higher-Order Singular Value Decomposition (HOSVD) method to decompose the tensor data. Set the ranks after decomposition to [10, 10, 3], that is, the first 10, the first 10, and the first 3 singular values are retained for the time, frequency, and feature dimensions respectively. The core tensor and factor matrices obtained from the decomposition are used to calculate the pulse dynamic indicators. For example, the sum of the singular values in the time dimension represents the stability of the pulse signal over time, and the distribution of the singular values in the frequency dimension reflects the energy distribution characteristics of the pulse. The calculation results of the dynamic indicators are saved as an Excel file named pulse_dynamic_indicators.xlsx. Import the pulse dynamic indicator time series table into the Python environment and use the NumPy and Matplotlib libraries to perform multi-lead signal synchronization processing. Load the pulse_dynamic_indicators.xlsx file, assuming the data contains the pulse signals of 3 leads. Use the Dynamic Time Warping (DTW) algorithm to calculate the optimal time alignment path between different lead signals. For example, for lead 1 and lead 2, the calculated alignment path shows that lead 2 is delayed by 5 sampling points relative to lead 1. According to the alignment result, calculate the phase difference and generate a multi-lead pulse phase difference feature map. The map shows the phase difference distribution between different leads in the form of a heat map and is saved as a PNG image file named phase_difference_map.png. In MATLAB, use the Image Processing Toolbox to analyze the phase_difference_map.png. Load the image file and convert it to a grayscale image. Use the graythresh function to calculate the global threshold of the image and binarize the image to highlight the main feature regions. Perform morphological processing (such as dilation and erosion operations) on the binarized image to extract the morphological features related to blood vessel elasticity. Use the pre-established empirical formula between the phase difference and blood vessel elasticity parameters (the formula is based on clinical experimental data) to calculate the blood vessel elasticity index sequence. For example, within a certain time window, the blood vessel elasticity modulus corresponding to the phase difference is 15 kPa. The calculation results are saved as a text file named artery_elasticity.txt. Use the PyTorch library in Python to model the spatio-temporal evolution of the blood vessel elasticity index sequence.Load the artery_elasticity.txt file and convert the data into a tensor format. Construct a one-dimensional convolutional neural network (CNN) model with a convolutional kernel size of 3, a stride of 1, and 16 output channels. The input of the model is the time series of vascular elasticity indicators, and the output is the feature map extracted by convolution. By training the model, the features reflecting the spatio-temporal changes of vascular elasticity are extracted. Combine these features with the previous pulse dynamic indicators to generate a pulse time series indicator sequence, which includes timestamps and comprehensive pulse indicator values.

[0036] Through zero-drift correction, the present invention effectively eliminates the baseline drift problem in pulse parameters and improves the data stability. Through the identification of the main wave morphological features of the pulse and subsequent time-varying morphological clustering analysis, not only the key morphological features of the pulse waveform are captured, but also its variation law over time is revealed. By constructing the joint time-frequency feature tensor data of the pulse, the time domain and frequency domain information are further integrated, providing a more comprehensive feature basis for subsequent analysis. Through the calculation of pulse dynamic indicators and the synchronous processing of multi-lead signals, the dynamic characteristics of the pulse and their phase differences can be quantified, thus providing a reliable basis for the calculation of vascular elasticity parameters. The finally generated pulse time series indicator sequence integrates the pulse dynamic characteristics and vascular elasticity information, comprehensively reflecting the dynamic change process of the pulse, and providing more detailed, quantitative and dynamic feature data for the pulse diagnosis analysis in traditional Chinese pediatrics.

[0037] Preferably, generating the body surface temperature rhythm indicator sequence based on the body region temperature feature table includes: Perform dynamic baseline correction on the body region temperature feature table to obtain a corrected body surface temperature time series table; Perform spatio-temporal interpolation on the corrected body surface temperature time series table to generate a body surface temperature spatio-temporally continuous distribution field; Extract body surface thermal characteristics from the body surface temperature spatio-temporally continuous distribution field to generate a body surface heat conduction parameter sequence; Construct a temperature feature attractor graph based on the body surface heat conduction parameter sequence; Perform body surface temperature complexity analysis based on the temperature feature attractor graph to generate a body surface temperature complexity indicator table; Perform oscillation mode decomposition on the body surface temperature complexity indicator table to obtain a temperature oscillation feature set; Perform wavelet coherence analysis on the temperature oscillation feature set to generate a body surface-visceral temperature intensity time series graph; Generate the body surface temperature rhythm indicator sequence based on the body surface-visceral temperature intensity time series graph.

[0038] In this embodiment, the Thermal Analysis Pro software is used to perform dynamic baseline correction on the body area temperature characteristic table. This characteristic table contains the body surface temperature data of the patient during the detection process, covering multiple areas such as the head, neck, trunk, and limbs. The software reads the original temperature data file, which records the temperature changes in each area during the detection period. In the software, select the dynamic baseline correction function module, set the time window length to 5 minutes, and the moving step size to 1 minute. The software calculates the median temperature within each time window as the baseline reference value, and takes the difference between the original temperature data and the baseline reference value as the corrected temperature value. For example, in the trunk area, the median temperature of the first 5-minute window is 36.2 °C, and the original temperature at a certain time point is 36.5 °C, then the corrected temperature is 0.3 °C. After processing, a corrected body surface temperature time series table is obtained. Import the corrected body surface temperature time series table into the Spatial Interpolation Toolbox software. This software is a spatio-temporal interpolation tool widely used in the biomedical field. In the software, select the Kriging interpolation method as the interpolation algorithm, and set the interpolation grid spacing to 2 cm × 2 cm. The software generates a spatio-temporal continuous distribution field of the body surface temperature based on the corrected temperature data and the geometric position information of each body area. For example, in the trunk area, the software calculates the temperature value of each grid point based on the data collected by adjacent temperature sensors, forming a smooth temperature distribution surface. In the Spatial Interpolation Toolbox software, use the thermal feature extraction function module to analyze the spatio-temporal continuous distribution field of the body surface temperature. The software automatically calculates thermal characteristic parameters such as thermal conductivity, thermal diffusivity, and heat capacity. For example, for the temperature distribution field in the trunk area, the software calculates the thermal conductivity as 0.2 W / (m·K), the thermal diffusivity as 0.1 cm² / s, and the heat capacity as 3.5 J / (g·K). These parameters reflect the heat conduction characteristics of different areas of the body surface. The thermal characteristic parameter sequence is saved in the CSV file format, which contains timestamps, area identifiers, and corresponding thermal characteristic values. Import the body surface thermal characteristic parameter sequence into the Chaos Analysis Suite software. In the software, select the phase space reconstruction function module, and set the delay time to 3 and the embedding dimension to 4. The software constructs a temperature characteristic attractor graph based on the thermal characteristic parameter sequence. For example, in the time series of thermal conductivity in the trunk area, the software generates a three-dimensional attractor graph through phase space reconstruction, showing the trajectory of thermal conductivity in the phase space. The graph uses different colors to represent the density of the trajectory, intuitively demonstrating the internal law of body surface temperature changes. In the Chaos Analysis Suite software, use the complexity analysis function module to deeply analyze the temperature characteristic attractor graph. The software calculates the fractal dimension and Lyapunov exponent of the attractor to evaluate the complexity of body surface temperature changes.For example, the fractal dimension of the temperature feature attractor in the trunk area is 2.3, and the Lyapunov exponent is 0.15, indicating that the temperature change in this area has a medium degree of complexity and chaotic characteristics. The complexity analysis results are saved as an Excel file, which contains the fractal dimension and Lyapunov exponent indicators of each area. The body surface temperature complexity index table is imported into the Vibration Analysis Toolbox software. In the software, the empirical mode decomposition (EMD) method is selected to decompose the body surface temperature time series data into multiple intrinsic mode functions (IMFs). For example, the temperature time series data in the neck area is decomposed into 5 IMF components and a residual component. Each IMF component represents the temperature oscillation mode in a different frequency band. The software extracts the oscillation frequency, amplitude, and energy characteristics of each IMF component to form a temperature oscillation feature set. The temperature oscillation feature set and the visceral temperature reference data (assumed to be obtained from an esophageal temperature sensor) are imported into the Wavelet Coherence Analyzer software. In the software, the Morlet wavelet is selected as the analysis mother wavelet, and the scale range is set to 1 - 30 days with a step size of 0.5 days. The software performs wavelet coherence analysis on the body surface temperature and visceral temperature data to generate a body surface-visceral temperature intensity time series graph. For example, the software shows that during the 5th - 10th day of detection, there is a significant coherence between the body surface temperature and the visceral temperature within the 2 - 5 day cycle range, and the coherence coefficient reaches above 0.7. According to the time series graph, the software extracts the coherence intensity and phase difference characteristics to generate a body surface temperature rhythm index sequence, which contains timestamp, coherence intensity, and phase difference indicators.

[0039] The present invention effectively eliminates the baseline drift in the body surface temperature data through dynamic baseline correction, improving the stability and accuracy of the data. Through spatio-temporal interpolation processing, a continuous distribution field of the body surface temperature is further generated, providing a more complete and smooth data basis for subsequent thermal feature extraction. Through the extraction of body surface thermal features and the construction of the temperature feature attractor graph, not only the conduction characteristics of the body surface temperature are captured, but also the internal laws and complexities of the temperature change are revealed. The body surface temperature complexity index table generated through the complexity analysis based on the attractor graph further quantifies the complexity characteristics of the body surface temperature change. Through the oscillation mode decomposition and wavelet coherence analysis, the periodic characteristics of the body surface temperature and its correlation with the visceral temperature are deeply explored, and the generated body surface-visceral temperature intensity time series graph intuitively reflects the dynamic relationship between the body surface and visceral temperatures. The finally generated body surface temperature rhythm index sequence synthesizes the complexity and oscillation characteristics of the body surface temperature, comprehensively reflecting the rhythmic changes of the body surface temperature over time.

[0040] Especially importantly, the construction of the pediatric data visual coding rules based on the tongue image time series evolution index sequence, the pulse condition evolution time series index sequence, and the body surface temperature rhythm index sequence includes: Construct visual coding rules for pediatric data based on the sequence of tongue image temporal evolution indicators, the sequence of pulse condition evolution temporal indicators, and the sequence of body surface temperature rhythm indicators; Construct a physiological index evolution trajectory map based on the sequence of tongue image temporal evolution indicators, the sequence of pulse condition evolution temporal indicators, and the sequence of body surface temperature rhythm indicators; Construct a basic framework for temporal sector navigation based on the physiological index evolution trajectory map; Design a multi-modal data interface for the basic framework of temporal sector navigation to obtain a unified temporal indexing architecture; Design an interactive trigger event stream based on the unified temporal indexing architecture to obtain a user intention recognition framework; Construct a view linkage controller according to the user intention recognition framework to obtain a multi-view interaction feedback mechanism; Perform focus data enhancement and context data reduction on the multi-view interaction feedback mechanism to obtain a focus-context visual enhancement scheme; Perform dynamic detection of time critical nodes based on the focus-context visual enhancement scheme to obtain a highlighting scheme for critical time nodes; Perform visual coding design on the highlighting scheme for critical time nodes to obtain visual coding rules for pediatric data.

[0041] In this embodiment, the Medical Design Studio software is used to construct the visual coding rules for pediatric data. First, load the sequence files of tongue image temporal evolution index, pulse condition evolution temporal index, and body surface temperature rhythm index. According to the established visualization requirements, set the tongue image index to be represented by red, the pulse condition index to be represented by blue, and the temperature index to be represented by yellow. In the software, set the data range: the tongue image index range is 0 - 100, the pulse condition index range is 0 - 200, and the temperature index range is 34 - 38 °C. According to these ranges, the software automatically generates a color mapping rule to convert the data value into the corresponding color shade. For example, the tongue image value of 50 corresponds to medium red, the pulse condition value of 100 corresponds to medium blue, and the temperature of 36 °C corresponds to medium yellow. Finally, the constructed visual coding rules for pediatric data are saved in a file, which defines the visualization representation of different types of data. Import the previously constructed visual coding rule file into the Trajectorymapper software. This software is a trajectory atlas generation tool. In the software, select a three-dimensional visualization mode with time as the horizontal axis and different physiological indicators as the vertical axis. Set the time axis range to the total duration from the start to the end of the detection. The software draws the physiological indicator evolution trajectory atlas according to the visual coding rules and the physiological indicator data. For example, the trajectory line of the tongue image index is displayed in red, and as the tongue image value changes, the color shade of the trajectory line changes accordingly; the pulse condition index is displayed as a blue trajectory line, and the color shade is also adjusted according to the numerical change. The atlas is displayed in the form of an interactive three-dimensional model, supporting operations such as rotation and scaling. The constructed trajectory atlas is saved as a file and stored in the hospital's visualization resource library. In the Trajectorymapper software, use the sector navigation plug-in to construct the basic framework of the temporal sector navigation. Set the center point of the sector to correspond to the start time of the detection, and the radius of the sector represents the time progress. From the center outwards are key time nodes such as the 1st day, 3rd day, 7th day, 14th day, etc. Each time node is represented by a small sector area, and the angle size of the sector is allocated according to the time interval. The total angle of the entire sector navigation is set to 270 degrees. The software projects the physiological indicator evolution trajectory atlas onto the sector navigation, and the status of each physiological indicator is represented by small dots of different colors. For example, at the position of the 3rd day, the red dot represents the tongue image status, and the blue dot represents the pulse condition status. The initially constructed navigation framework is saved as the fan_framework.fnav file. In the Fan Framework Designer software, perform a multi-modal data interface design on the basic framework of the temporal sector navigation. Select to add a floating prompt function on the sector area. When the mouse hovers over the small sector area of a certain time point, a floating window pops up to display the detailed physiological indicator data of that time point. Set the data loading delay to 300 milliseconds to avoid interface jamming caused by frequent operations.Meanwhile, integrate the database connection settings and connect to the hospital's PediatricData2025 database to load the data corresponding to the time points from the Tongue, Pulse, and Temperature tables. After completing the interface design, the software automatically generates a unified time-series index architecture, using the timestamp as the unified index key to associate various modal data. The final saved architecture file is UnifiedIndex.ui. Design the interaction trigger event flow in the Interaction Flow Architect software. On the time-series fan-shaped navigation, set the single-click event to view detailed data and the double-click event to compare data at adjacent time points. Add the keyboard shortcut Ctrl+D for quick comparison mode switching. The software records the user operation frequency. When the single-click operation frequency exceeds 5 times per minute, trigger the intelligent recommendation function to display relevant data of interest. The user intention recognition framework contains detailed definitions of event types, trigger conditions, and response actions. For example, when the user frequently switches views between the 7th day and the 14th day positions, the system automatically recommends and displays the trend analysis data within this time range to improve the user operation efficiency. In the Multi View Linker software, construct the view linkage controller. Set the main view as the fan-shaped navigation, and the auxiliary views as the detailed data table and the trend line chart. When the user selects a certain time point on the fan-shaped navigation, the detailed data table automatically updates to display the specific values of tongue image, pulse condition, and temperature at that time point; the trend line chart highlights the position of the selected time point and displays the change trend lines of adjacent time points. Add a smooth transition animation with a 0.5-second fade effect when switching views to enhance visual coherence. The constructed multi-view interaction feedback mechanism is saved as the ViewLink.vlc file. For example, after the user clicks on the 10th day position, the detailed data table smoothly transitions within 0.5 seconds to display the data for the 10th day, and at the same time, the highlighted part of the trend line chart is also updated synchronously, providing a more intuitive display of data association for the user. In the FocusContext Enhancer software, perform focus data enhancement and context data reduction. On the fan-shaped navigation, set the currently selected time point as the focus area, magnify its display, and display the non-focus area with 50% transparency to guide the user to focus on the focus data. In the detailed data table, the focus data rows are highlighted by using bold font and increased row height. For the context data, only display a simplified view of the key indicators, such as only showing the integer part of the temperature, and the main trend labels of tongue image and pulse condition. After the enhancement scheme is saved as the FocusScheme.fcs file and displayed in the interface, the focus area is automatically magnified and highlighted, and the context data is presented in a simplified form, enhancing the interface information hierarchy and helping the user quickly locate key information. In the Key Pointmarker software, perform dynamic detection of time key nodes based on the focus-context visual enhancement scheme.Set a key node detection algorithm. When the physiological indicators change beyond a preset threshold (such as the tongue image change being greater than 20%, the pulse condition change being greater than 30%, and the temperature change being greater than 1°C), it is automatically marked as a key time node. On the fan-shaped navigation, the key nodes are displayed as prominent star marks, and the mark colors correspond to the visual coding rules according to the indicator types. For example, the key nodes of tongue images are marked with red stars, the pulse conditions with blue stars, and the temperatures with yellow stars. After the plan is saved as a KeyPoints.kps file, the key time nodes are automatically highlighted as star marks on the fan-shaped navigation, guiding users to quickly identify the important change moments in the data and improving the data interpretation efficiency. In the Medical Design Studio software, perform the final visual coding design for the key time node highlighting scheme. Dynamically adjust the size of the star marks according to the key degree. The key nodes with a change amplitude exceeding 150% of the threshold are displayed as large stars, those with a change amplitude between 100% - 150% are displayed as medium stars, and those with a change amplitude between the preset threshold and 100% are displayed as small stars. At the same time, add a hover hint function. When the mouse hovers over the star mark, display the detailed information of the key node, including the type of indicator that has changed, the change amplitude, and the clinical significance. The finally designed visual coding rules for pediatric data integrate all visualization elements such as color coding, node marking, and interactive hints.

[0042] Especially importantly, the visualization of the tongue image time-series evolution indicator sequence, the pulse condition evolution time-series indicator sequence, and the body surface temperature rhythm indicator sequence according to the pediatric data visual coding rules includes: Visualize the tongue image time-series evolution indicator sequence, the pulse condition evolution time-series indicator sequence, and the body surface temperature rhythm indicator sequence according to the pediatric data visual coding rules; Construct a hierarchical detail display mechanism according to the pediatric data visual coding rules to obtain a pediatric data detail display framework; Establish an inter-modal navigation channel based on the pediatric data detail display framework to obtain an inter-modal seamless connection navigation unit; Draw a dynamic symptom association force-directed graph based on the physiological indicator evolution trajectory atlas and the inter-modal seamless connection navigation unit to obtain a dynamic symptom association force-directed graph; Map the dynamic symptom association force-directed graph to a human three-dimensional model to obtain a semantic-enhanced symptom interaction relationship model; Construct a multi-angle comparison view framework based on the semantic-enhanced symptom interaction relationship model to obtain a pediatric data spatio-temporal comparison analysis framework; Formulate a symptom evolution visualization component according to the pediatric data spatio-temporal comparison analysis framework to obtain a symptom intensity spatio-temporal distribution map; Visualize the evolution of syndrome types by modeling the spatio-temporal distribution map of symptom intensity, and obtain a dynamic visualization view of the evolution of syndrome types; visualize the spatio-temporal distribution map of symptom intensity and the dynamic visualization view of the evolution of syndrome types through a preset data display interface.

[0043] In this embodiment, the tongue image, pulse condition, and body surface temperature data are visualized using Tableau software. The sequence files of tongue image time-series evolution indicators, pulse condition evolution time-series indicators, and body surface temperature rhythm indicators are imported into Tableau. According to the previously designed visual coding rules for pediatric data, it is set that the tongue image indicators are represented in red, the pulse condition indicators are represented in blue, and the temperature indicators are represented in yellow. In Tableau, three line charts are created, corresponding to the tongue image, pulse condition, and temperature indicators respectively. For example, the line chart of the tongue image indicator shows the change trend of the tongue image value within the range of 0 - 100, the line chart of the pulse condition indicator shows the change trend of the pulse condition value within the range of 0 - 200, and the line chart of the temperature indicator shows the change trend of the temperature within the range of 34 - 38 °C. The finally generated visualization chart is saved as a Tableau workbook file. In the Tableau workbook, a framework for displaying the details of pediatric data is constructed. Using the hierarchical view function of Tableau, the data is displayed in layers according to the time hierarchy (hours, days, weeks). The default view is set to display the weekly-level data, and filters are added to the view to allow users to view the daily-level and hourly-level data through drill-down operations. For example, users can see the overall trends of the tongue image, pulse condition, and temperature in the weekly view. After clicking on the data of a certain week, the detailed data of each day within that week is automatically expanded and displayed. The constructed hierarchical detail display framework is saved in the workbook file to ensure that users can flexibly switch between different levels of details. In the Tableau workbook, a navigation channel between modalities is established using the action function. Three worksheets, namely tongue image details, pulse condition details, and temperature details, are set as navigation targets. When a user selects a data point at a certain time point in the main view (such as a red data point where the tongue image value exceeds 80), a filtering action for the tongue image details worksheet is triggered, and it automatically jumps to and displays the detailed tongue image information at that time point. At the same time, a button to return to the main view is set. The navigation channel settings are saved in the workbook file to ensure that users can smoothly navigate between different modality data and comprehensively understand the multi-dimensional information of pediatric data. In Gephi software, the physiological index evolution trajectory atlas file and the interaction log file of the seamless connection navigation unit between modalities are loaded. Nodes are set to represent the physiological index states at different time points, and edges represent the correlation strength between the indicators. The correlation strength is calculated based on the correlation coefficient between the indicators. Indicator pairs with a correlation coefficient greater than 0.6 are considered to have a strong correlation. For example, between time points with a high tongue image value (dark red) and a high temperature value (dark yellow), a thicker edge is drawn to represent a strong correlation. The software automatically layouts to generate a symptom correlation force-directed graph, showing the correlation network between different physiological indicators. The force-directed graph is saved in the native format of Gephi. In Blender software, the data of the symptom correlation force-directed graph is imported. The standard children's human model is loaded using the human model plugin of Blender.Map the nodes in the guidance diagram to the corresponding parts of the human body model. For example, map the tongue image nodes to the oral cavity position, the pulse condition nodes to the wrist position, and the temperature nodes to the surface of the torso. Set the size and color of the nodes to change dynamically according to the symptom intensity, and display the associated edges as semi-transparent lines. The finally generated semantic enhanced symptom interaction relationship model is saved in the Blender file format, supporting three-dimensional rotation and zoom viewing. In the 3dsmax software, load the semantic enhanced symptom interaction relationship model file. Utilize the multi-view function of 3dsmax to create four view windows, corresponding to the front view, side view, top view, and perspective view respectively. Set the camera parameters for each view. The front view shows the symptom distribution on the front of the human body, the side view shows the symptom distribution on the side, the top view shows the symptom distribution on the top, and the perspective view provides a three-dimensional perspective. For example, in the front view, the tongue image nodes are displayed at the oral cavity position, and the pulse condition nodes are displayed at the wrist position; in the perspective view, it can be clearly seen how the associated edges between the nodes are connected in three-dimensional space. The constructed multi-angle comparison view framework is saved as a 3dsmax scene file. In the Unity engine, create a spatio-temporal distribution map of symptom intensity. Utilize the Timeline function of Unity to set the time axis range to the total duration of the detection period. Map the symptom intensity data to a bar chart in three-dimensional space. The tongue image intensity is represented by a red bar chart, the pulse condition intensity is represented by a blue bar chart, and the temperature intensity is represented by a yellow bar chart. For example, at the 3rd day position on the time axis, the bar chart with a higher tongue image intensity value is displayed as a higher red bar. Through the animation system of Unity, produce a dynamic effect of the symptom intensity evolving over time. The finally generated spatio-temporal distribution map of symptom intensity is saved as a Unity scene file. In the Unity engine, load the spatio-temporal distribution map file of symptom intensity. Utilize the animation curve function of Unity to construct a syndrome type evolution model according to the classification criteria of traditional Chinese medicine syndrome types. Set the combined characteristics corresponding to the wind-heat syndrome as red tongue image, fast pulse condition, and high temperature, and the combined characteristics corresponding to the wind-cold syndrome as light tongue image, slow pulse condition, and low temperature. Through key frame animation technology, display the evolution process of the syndrome type in the time series. For example, show the characteristics of the wind-cold syndrome in the first 3 days of detection, and gradually transition to the characteristics of the wind-heat syndrome starting from the 4th day. The finally generated dynamic visualization view of syndrome type evolution is saved as a Unity prefab file, supporting reuse in different projects. In the Unity engine, create a comprehensive data display interface. Utilize the UI system of Unity to design an interface layout containing two main panels. The left panel is used to display the spatio-temporal distribution map of symptom intensity, and the right panel is used to display the dynamic visualization view of syndrome type evolution. Add a time slider controller to allow users to drag the slider on the bottom time axis to view the data at different time points. Set interactive buttons such as play, pause, and reset to facilitate user operation. The final visualization interface is packaged as a Unity executable file and deployed on the hospital workstation.

[0044] Therefore, in all aspects, the embodiments should be regarded as exemplary and non-limiting. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the application documents are intended to be encompassed within the present invention.

[0045] The above are only specific embodiments of the present invention, enabling those skilled in the art to understand or implement the present invention. Various modifications to these embodiments will be obvious to those skilled in the art. The general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to these embodiments shown herein, but rather to the broadest scope consistent with the principles and novel features invented herein.

Claims

1. A visualization data processing system applied to traditional Chinese medicine pediatrics, characterized in that It includes the following modules: A data acquisition module, which is used to acquire a pediatric physiological time series dataset; perform physiological event anchor point recognition on the pediatric physiological time series dataset to obtain a physiological event anchor point set; A time calibration module, which is used to perform noise suppression on the pediatric physiological time series dataset according to the physiological event anchor point set to obtain a noise-suppressed physiological time series dataset; Perform time calibration on the noise-suppressed physiological time series dataset to obtain a time-corrected physiological dataset; A feature extraction module, which is used to extract a tongue image set from the time-corrected physiological dataset; generate a tongue image feature vector set based on the tongue image set; Extract a pulse signal dataset from the time-corrected physiological dataset; construct a pulse characteristic parameter set based on the pulse signal dataset; extract a thermal imaging image sequence from the time-corrected physiological dataset; generate a body area temperature feature table based on the thermal imaging image sequence; An evolution index construction module, which is used to construct a tongue image time series index sequence based on the tongue image feature vector set; generate a pulse time series index sequence based on the pulse characteristic parameter set; generate a body surface temperature rhythm index sequence based on the body area temperature feature table; A visualization module, which is used to construct a visual coding rule for pediatric data based on the tongue image time series index sequence, the pulse time series index sequence, and the body surface temperature rhythm index sequence; visualize the tongue image time series index sequence, the pulse time series index sequence, and the body surface temperature rhythm index sequence according to the visual coding rule for pediatric data.

2. The application to the traditional Chinese medicine pediatric visualization data processing system according to claim 1, wherein, The acquisition of the pediatric physiological time series dataset includes: Performing multi-dimensional physiological information acquisition on children through multi-source traditional Chinese medicine pediatric detection devices to obtain a multi-modal traditional Chinese medicine pediatric physiological dataset; Performing device identification code registration on the multi-source traditional Chinese medicine pediatric detection devices to obtain a device identification code data table; using a preset network time protocol server to perform clock synchronization calibration on the multi-source traditional Chinese medicine pediatric detection devices to obtain a device reference time synchronization record; Performing data source parsing on the multi-modal traditional Chinese medicine pediatric physiological dataset according to the device identification code data table and the device reference time synchronization record to obtain a device data stream meta-information table; Extracting the timestamp field from the device data stream meta-information table and performing standardized conversion to obtain an original acquisition timestamp sequence; calculating the time deviation value between each detection device based on the original acquisition timestamp sequence to obtain a device time difference parameter table; Calculating the standard deviation of the time deviation of 5 consecutive measurements in the device time difference parameter table, and performing stability judgment according to the standard deviation of the time deviation and a preset stability threshold to obtain a time deviation stability judgment result; Performing reliability quantification on the device time difference parameter table according to the time deviation stability judgment result to obtain a time difference reliability evaluation result; performing weighted average correction on the device time difference parameter table according to the time difference reliability evaluation result to obtain a corrected device time difference parameter table; Performing dynamic time warping on the multi-modal traditional Chinese medicine pediatric physiological dataset based on the corrected device time difference parameter table to obtain a pediatric physiological time series dataset, where the pediatric physiological time series dataset includes pulse waveform data, tongue coating microscopic image sequences, and body surface thermal imaging data.

3. The application to the visualization data processing system for traditional Chinese medicine pediatrics according to claim 1, wherein Performing physiological event anchor point recognition on the pediatric physiological time series dataset includes: Performing real-time peak detection on the pulse waveform data to obtain a pulse peak position sequence; Calculate the time intervals between adjacent pulse wave peaks according to the pulse wave peak position sequence, and perform physiological validity screening on the pulse wave peak position sequence according to the time intervals between adjacent pulse wave peaks and the preset physiological cycle threshold range to obtain a standard pulse cycle marker point set; Perform frame-by-frame comparison on the tongue coating microscopic image sequence based on the inter-frame difference method to obtain the marker points of the rapidly changing tongue image frames; Calculate and mark the regional temperature gradient of the body surface thermal imaging data to obtain a temperature abrupt change region marker map; Based on the temperature abrupt change region marker map, count the time points of the temperature abrupt change regions in the consecutive frames of the body surface thermal imaging data to generate a thermal imaging event trigger time point sequence; Record the corresponding event trigger points in the pulse wave peak position sequence, the marker points of the rapidly changing tongue image frames, and the thermal imaging event trigger time point sequence as the initial physiological event anchor point set; Construct a cross-modal event time correspondence table based on the pulse wave peak position sequence, the marker points of the rapidly changing tongue image frames, and the thermal imaging event trigger time point sequence; Based on the cross-modal event time correspondence table, make the following judgments on each event anchor in the initial physiological event anchor set: If at least two modalities have trigger events within the ±200 ms window, mark the corresponding event anchor as a high-confidence anchor; otherwise, remove the corresponding event anchor, and record the high-confidence anchors as the physiological event anchor set.

4. The application to the visualization data processing system for traditional Chinese medicine pediatrics according to claim 1, characterized in that Time calibration of the noise-suppressed physiological time series dataset includes: Segment the noise-suppressed physiological time series dataset with a fixed window length of 10 seconds and an adjacent window overlap rate of 20% to generate a segmented time series dataset; Calculate the slope and intercept parameters of each segment of data in the segmented time series dataset to generate a segmented linear parameter table; Calculate the mean square residual of each segment of data in the segmented time series dataset according to the segmented linear parameter table, and perform fitting qualification determination and screening on the segmented linear parameter table according to the mean square residual and the preset fitting quality screening threshold to generate a valid segmented linear parameter table; Perform differential calculation on the slopes of adjacent segments in the valid segmented linear parameter table to generate an adjacent segment slope difference sequence; Identify the slope mutation points according to the adjacent segment slope difference sequence to generate a drift mode switching point sequence; Perform pattern matching on the valid segmented linear parameter table according to the preset typical drift mode and the drift mode switching point sequence to generate a drift mode classification table; Perform adaptive smoothing on the connection points of different mode junction types based on the drift mode classification table to generate a segmented drift characteristic curve, where the adaptive smoothing is specifically: When the junction type is an increasing / decreasing mode junction, use cubic Hermite interpolation for smoothing; When the junction type is an oscillating mode junction, use mean smoothing filtering for smoothing; Perform cubic spline interpolation on the segmented drift characteristic curve to generate a signal time drift characteristic curve; construct a time mapping rule according to the signal time drift characteristic curve; Perform time coordinate transformation on the noise-suppressed physiological time series dataset based on the time mapping rule to obtain a time-corrected physiological dataset, where the time-corrected physiological dataset includes calibrated pulse waveform data, calibrated tongue coating microscopic image sequence, and calibrated body surface thermal imaging data.

5. The application to the visual data processing system for traditional Chinese medicine pediatrics according to claim 1, characterized in that, Generating a tongue image feature vector set based on the tongue image set includes: Perform histogram equalization on the tongue image set to obtain an enhanced tongue image set; Perform semantic segmentation on the enhanced tongue image set to obtain the tongue body region mask map; perform region extraction on the enhanced tongue image set according to the tongue body region mask map to obtain the tongue body ROI image set; Perform color space conversion and decomposition on the tongue body ROI image set to obtain the tongue image feature channel group; Perform wavelet transform on the tongue image feature channel group to obtain the tongue image texture wavelet coefficient table; perform statistical feature extraction on the tongue image texture wavelet coefficient table to obtain the tongue image texture feature table; Calculate the local binary pattern feature based on the tongue body ROI image set to obtain the tongue image LBP feature map; perform color distribution feature analysis according to the tongue body ROI image set to obtain the tongue color feature vector; Perform feature fusion on the tongue image texture feature table, the tongue image LBP feature map and the tongue color feature vector to obtain the tongue image feature vector set.

6. The application to the visualization data processing system for traditional Chinese medicine pediatrics according to claim 1, wherein, Construct the pulse characteristics parameter set based on the pulse signal data set, including: Perform pulse period detection and segmentation on the pulse signal data set to obtain the single-period pulse segment set; Perform continuous wavelet transform on the single-period pulse segment set to obtain the pulse time-frequency spectrum map set; Extract the frequency band energy feature based on the pulse time-frequency spectrum map set to obtain the pulse frequency domain feature set; Calculate the time-domain characteristic parameters of the noise-reduced pulse waveform data to obtain the pulse time-domain characteristic table; Construct the pulse characteristics parameter set according to the pulse frequency domain feature set and the pulse time-domain characteristic table.

7. The application to the visualization data processing system for traditional Chinese medicine pediatrics according to claim 1, wherein Generate the body region temperature feature table based on the thermal imaging image sequence, including: Perform temperature correction and pseudo-color mapping on the thermal imaging image sequence to obtain the standard body surface thermal imaging map set; Perform threshold segmentation based on the standard body surface thermal imaging map set to obtain the body surface region temperature mask map; Perform region division and marking on the body surface region temperature mask map to obtain the body part temperature region map; Calculate the temperature statistical features of each region based on the body part temperature region map to obtain the body region temperature feature table.

8. The application to the traditional Chinese medicine pediatrics visualization data processing system according to claim 1, wherein Construct the tongue image time series index sequence based on the tongue image feature vector set, including: Perform color correction on the tongue image feature vector set to obtain the tongue image color correction feature vector set; Extract the boundary feature between the tongue body and the tongue coating based on the tongue image color correction feature vector set to obtain the tongue body-tongue coating boundary contour sequence; Perform tongue image contour evolution simulation based on the tongue body-tongue coating boundary contour sequence to obtain the tongue image contour evolution trajectory data; Calculate the tongue body stretching and shrinking dynamic parameters based on the tongue image contour evolution trajectory data to obtain the tongue body movement index set; Perform wavelet packet decomposition on the tongue image color correction feature vector set to obtain the tongue image texture wavelet energy time series spectrum; Perform anisotropy analysis of the tongue coating roughness on the tongue image texture wavelet energy time series spectrum to obtain the tongue coating roughness-direction correlation index table; Calculate the regional humidity index of the tongue body-tongue coating boundary contour sequence to obtain the tongue body humidity index time series sequence; Generate the tongue image time series index sequence based on the tongue body movement index set, the tongue coating roughness-direction correlation index table and the tongue body humidity index time series sequence.

9. The application to the visualization data processing system for traditional Chinese medicine pediatrics according to claim 1, characterized in that, Generate the pulse time series index sequence based on the pulse characteristics parameter set, including: Perform zero drift correction on the pulse characteristics parameter set to obtain the corrected pulse parameter time series sequence; Perform pulse main wave morphology feature recognition based on the corrected pulse parameter time series sequence to generate the pulse main wave morphology time series feature sequence; Perform time-varying morphological clustering on the time series feature sequence of the main wave form of the pulse condition, and generate a time-varying morphological clustering label set for the pulse wave; Construct a time-frequency joint feature tensor data for the pulse condition based on the time-varying morphological clustering label set of the pulse wave; Calculate the pulse condition dynamic index for the time-frequency joint feature tensor data of the pulse condition, and generate a time series table of the pulse condition dynamic index; Perform multi-lead signal synchronization according to the time series table of the pulse condition dynamic index, and generate a multi-lead pulse condition phase difference feature map; Calculate the vascular elasticity parameter based on the multi-lead pulse condition phase difference feature map, and generate a vascular elasticity index sequence; Perform spatio-temporal evolution modeling on the vascular elasticity index sequence, and generate a time series index sequence of the pulse condition.

10. The application to the traditional Chinese medicine pediatric visualization data processing system according to claim 1, characterized in that Generate a body surface temperature rhythm index sequence based on the body region temperature feature table, including: Perform dynamic baseline correction on the body region temperature feature table to obtain a corrected body surface temperature time series table; Perform spatio-temporal interpolation on the corrected body surface temperature time series table to generate a spatio-temporally continuous distribution field of the body surface temperature; Extract the body surface thermal characteristics from the spatio-temporally continuous distribution field of the body surface temperature, and generate a body surface heat conduction parameter sequence; Construct a temperature feature attractor map according to the body surface heat conduction parameter sequence; Perform body surface temperature complexity analysis based on the temperature feature attractor map, and generate a body surface temperature complexity index table; Perform oscillation mode decomposition on the body surface temperature complexity index table to obtain a temperature oscillation feature set; Perform wavelet coherence analysis on the temperature oscillation feature set to generate a body surface-visceral temperature intensity time series graph; generate a body surface temperature rhythm index sequence based on the body surface-visceral temperature intensity time series graph.

Images