Attaching and pricking effect real-time monitoring system for thyroid cancer postoperative neck and shoulder function rehabilitation

Abstract

The invention relates to the technical field of intelligent rehabilitation monitoring, in particular to a real-time sticking and pricking effect monitoring system for thyroid cancer postoperative neck and shoulder function rehabilitation, which comprises a vector sequence generation module, a convergence section construction module, a direction difference calculation module, a stable section screening module and an effect monitoring module. According to the method, continuous image frames are collected, a grid observation node shadow direction vector sequence is extracted, the shadow flow trend caused by skin surface microscopic deformation is captured, and the relative deviation between the static texture main direction and the dynamic shadow vector angle is used for representing the real-time coupling degree of the sticking structure and the skin texture. Stable continuous time slices are screened through differential variation amplitude to eliminate invalid vibration interference, numerical variance is counted and calculated based on direction difference sequence dispersion degree, dynamic stability depth quantitative evaluation of a target area is achieved, and an accurate quantitative index reflecting muscle skin movement coordination is provided. The defect that the micromechanics interaction state is difficult to perceive through macroscopic contour monitoring is overcome.

Description

Technical Field

[0001] This invention relates to the field of intelligent rehabilitation monitoring technology, and in particular to a real-time monitoring system for the effect of taping on neck and shoulder function rehabilitation after thyroid cancer surgery. Background Technology

[0002] The field of intelligent rehabilitation monitoring technology encompasses technologies that monitor human movement status, biomechanical characteristics, and local tissue changes through methods such as sensor acquisition, image analysis, posture recognition, continuous data recording, and computational processing. It typically includes signal acquisition, image or motion data extraction, time series feature comparison, and processing to generate rehabilitation assessment information based on the measured data, in order to achieve continuous observation of the activity of body parts and the state of the attached physical medium during rehabilitation training.

[0003] Among them, the real-time monitoring system for the taping effect of thyroid cancer surgery neck and shoulder function rehabilitation refers to a technical solution used to record the deformation of the taping area, the movement trajectory of the skin surface and the relative displacement of the attached structure during neck and shoulder training. It usually acquires training images through fixed camera equipment, identifies the neck and shoulder area through image segmentation, judges the changes in the taping contour through edge tracking, determines the local stretching status through pixel displacement comparison, and records the neck and shoulder activity path through motion trajectory extraction.

[0004] Existing technologies rely on fixed recording equipment and determine contour changes based on edge tracking. They only capture macroscopic displacement and deformation in the neck and shoulder area, ignoring the microscopic physiological tremors and subtle tension fluctuations that accompany the skin surface after surgery. Simply using pixel displacement comparison makes it difficult to distinguish between the substantial movement of muscle traction and the ineffective displacement of epidermal sliding. Based solely on external contour features, it is impossible to perceive the mechanical coupling state between the tying structure and skin texture under dynamic interaction. The lack of analysis of high-frequency light and shadow flow details makes it difficult to accurately capture the subtle differences in the coordination of deep muscle and skin movements, resulting in insufficient accuracy in assessing the postoperative functional recovery status. Summary of the Invention

[0005] To address the technical problems existing in the prior art, this invention provides a real-time monitoring system for taping effects in neck and shoulder function rehabilitation after thyroid cancer surgery. The technical solution is as follows:

[0006] On the one hand, a real-time monitoring system for the taping effect in the rehabilitation of neck and shoulder function after thyroid cancer surgery is provided. This system includes: The vector sequence generation module acquires continuous image frames of the neck and shoulder after thyroid cancer surgery, divides the single frame image into a grid, extracts the image pixel points at the center of the grid as observation nodes, and determines the light and shadow direction vector sequence of adjacent observation nodes. The convergence segment construction module selects a specified observation node as the center position, calculates the angular deviation between the light and shadow direction vector sequences corresponding to the adjacent observation nodes in the four spatial neighborhoods of the specified observation node, and marks the light and shadow direction convergence segment. The orientation difference calculation module scans the brightness gradient within the grid of the first image observation node in a series of post-thyroid cancer surgery neck and shoulder images, compares it with the light and shadow orientation vector sequence, and filters the light and shadow-texture orientation difference sequence. The stable segment selection module determines the variation amplitude of the directional difference between adjacent time intervals in the light and shadow-texture directional difference sequence and selects stable light and shadow-texture coupling segments. The effect monitoring module determines the coordinates of the target domain for bonding based on the spatial coordinate indices of the observation nodes marked as the convergence zone of the light and shadow direction and the stable zone of the light and shadow-texture coupling, quantifies the dynamic stability of the bonding area, and obtains the bonding effect monitoring results.

[0007] As a further embodiment of the present invention, the light and shadow direction vector sequence includes a time step number, a node position identifier, and a light and shadow vector angle value. The light and shadow direction convergence segment specifically includes a convergence grid region index, a four-neighbor angle deviation set, and a convergence judgment threshold marker. The light and shadow-texture direction difference sequence includes a texture gradient principal direction value, a texture flow field relative angle, and a time series record list. The light and shadow-texture coupling stable segment includes a first-order difference variation amplitude, a relative static time window, and a stable feature grid location. The bonding effect monitoring result specifically refers to the boundary coordinates of the aggregation region, the sequence numerical variance term, and the bonding effect monitoring index.

[0008] As a further aspect of the present invention, the vector sequence generation module includes: The node pixel extraction submodule acquires continuous image frames of the neck and shoulder after thyroid cancer surgery. It performs grid segmentation on each single frame of the continuous image frames of the neck and shoulder after thyroid cancer surgery, locates the center position of the grid and extracts the pixel point at the center position of the grid as a spatially fixed observation node, reads the gray value of each observation node on the single frame image, and generates the center pixel set of the observation node. The brightness gradient comparison submodule extracts the brightness values ​​of the observation nodes in the central pixel set of the observation node between temporally adjacent single-frame images, performs temporal difference processing, calculates the brightness change rate of the current observation node, performs gradient comparison with the brightness change rate of spatially adjacent observation nodes, determines the difference magnitude and positive / negative relationship between the two, and obtains the node brightness gradient comparison result. The vector sequence generation submodule determines the light and shadow vector angle representing the light and shadow flow trend on the skin surface at each observation node at the current moment based on the node brightness gradient comparison results, and arranges the light and shadow vector angles in each consecutive frame image in chronological order to generate a light and shadow direction vector sequence.

[0009] As a further aspect of the present invention, the convergence segment construction module includes: The deviation value calculation submodule selects a specified observation node as the center position and traverses the light and shadow direction vector sequence corresponding to the adjacent observation nodes in the four neighboring spatial domains of the specified observation node. For each time point in the sequence, it reads the light and shadow vector angle values ​​of the center position and the four neighboring positions respectively, calculates the angle deviation between the light and shadow vector angle of the observation node at the center position and the light and shadow vector angle of the observation nodes at the four neighboring positions, and generates a set of local direction angles. The threshold interval determination submodule compares each angle deviation in the set of local direction angles with a preset convergence angle threshold, counts the frequency of each adjacent observation node's deviation value falling within the convergence angle threshold range throughout the entire time period of the sequence, and filters the observation nodes whose corresponding frequency occupies the majority in the total time points to obtain the threshold node index. The convergence segment marking submodule locates the grid space region corresponding to the observation node that meets the threshold node index in the pixel grid coordinate system, identifies the adjacency relationship of the grid space region on the image plane, merges adjacent grid regions with the same flow direction in a spatial manner, determines the geometric coverage of the merged region, and marks it as an attribute region with the same flow direction in space to obtain the light and shadow direction convergence segment.

[0010] As a further aspect of the present invention, the direction difference calculation module includes: The texture direction extraction submodule selects the first frame image in the continuous image frames of the neck and shoulder after thyroid cancer surgery, performs brightness gradient scanning on all pixels in the grid area where the observation node is located in the first frame image, detects the gradient magnitude distribution in each direction in the area, and extracts the direction with the maximum gradient magnitude as a static index characterizing the inherent skin features of the area, which is then identified as the main direction of skin texture. The relative angle calculation submodule takes the static main direction of the skin texture as a reference, traverses and reads the light and shadow vector angle data corresponding to each time point in the light and shadow direction vector sequence, calculates the relative angle value between the static main direction of the texture and the dynamic light and shadow vector angle, and generates a set of relative angle values. The difference sequence filtering submodule arranges each relative angle value in the relative angle value set according to the time step order, records the change trajectory of the relative angle value in the time dimension, reflects the coupling relationship between the texture main axis and the light and shadow flow direction of the skin surface, and generates a light and shadow-texture direction difference sequence.

[0011] As a further aspect of the present invention, the process reflecting the coupling relationship between the texture principal axis and the direction of light and shadow flow on the skin surface specifically includes: Check the range of values ​​for each relative angle value in the set of relative angle values. If the relative angle value exceeds 90 degrees, perform angle complementation operation to convert it into an acute angle value within 90 degrees, and uniformly quantify the degree of deviation between the texture main axis and the light and shadow vector. Create a time series container indexed by time step, and store the converted acute angle values ​​into the time series container in the original time order in the light and shadow direction vector sequence. The generated time series container records the trajectory data of the deviation degree dynamically fluctuating over time. This numerical sequence represents the dynamic coupling state between the texture axis and the direction of light and shadow flow on the skin surface in the continuous time domain.

[0012] As a further aspect of the present invention, the stable segment screening module includes: The variation amplitude calculation submodule locates the direction difference values ​​between adjacent time points in the light and shadow-texture direction difference sequence, performs first-order difference calculation, obtains the absolute value of the difference between the direction difference values ​​between adjacent time points, arranges them in the original time step order, and generates the variation amplitude of adjacent direction differences. The continuous segment recognition submodule introduces a preset stability judgment threshold as a baseline for measuring data fluctuations. It compares the variation amplitude of each adjacent direction difference with the stability judgment threshold, filters data points that are continuously less than the stability judgment threshold, extracts continuous time intervals of data points on the time axis, and confirms that the texture and the light and shadow flow on the skin surface remain relatively static within the time interval to obtain stable continuous time segments. The stable segment marking submodule retrieves the observation nodes corresponding to the stable continuous time segment, locates the geometric grid position of the target observation node in the spatial grid coordinate system, confirms the stability attributes of multiple geometric grid positions in the process of light and shadow and texture interaction, and generates a stable light and shadow-texture coupling segment.

[0013] As a further aspect of the present invention, the process of confirming the stability properties of multiple geometric mesh positions during the interaction between light and shadow and texture is specifically as follows: The total time span of all time segments determined to be stable and continuous at the same geometric grid location is calculated. Divide the total time span by the total time period length covered by the light and shadow-texture direction difference sequence to calculate the stability coverage ratio of the geometric mesh position in the time dimension; The stability coverage ratio is compared with the preset steady-state judgment benchmark ratio. When the stability coverage ratio is greater than or equal to the steady-state determination benchmark ratio, it is confirmed that the geometric mesh position remains stable under the interaction of light and shadow flow and texture features.

[0014] As a further aspect of the present invention, the effect monitoring module includes: The target domain coordinate determination submodule obtains the spatial coordinate indexes of the observation nodes marked as the convergence section of the light and shadow direction and the spatial coordinate indexes of the observation nodes marked as the stable section of light and shadow-texture coupling. It performs an intersection filtering operation on the two sets of coordinate indices to determine overlapping candidate points. Based on the spatial adjacency distance of the overlapping candidate points in the grid array, it performs connectivity analysis and aggregates them to form a connected region, generating the target domain coordinates for bonding. The stability quantization submodule iterates through the observation nodes that fall within the coordinate range of the patching target domain, extracts the light-shadow-texture direction difference sequence corresponding to each observation node and performs discreteness statistics, calculates the numerical variance of the sequence in the time dimension, quantifies the fluctuation of the light-shadow and texture coupling state in the patching area, and obtains the dynamic stability value of the region. The monitoring results generation submodule introduces a preset rehabilitation training standard threshold as a benchmark parameter, compares the dynamic stability value of the region with the rehabilitation training standard threshold, calculates the degree of deviation between the two values, assesses the recovery status of postoperative neck and shoulder function based on the degree of deviation, outputs a quantitative index reflecting the coordination of neck and shoulder muscle and skin movement, and generates taping effect monitoring results.

[0015] As a further aspect of the present invention, the process of performing connectivity analysis and aggregating overlapping candidate points to form connected regions based on their spatial adjacency distance in the grid array is as follows: Calculate the Euclidean distance between the geometric centers of any two overlapping candidate points in the grid array; The Euclidean distance of the geometric center is numerically compared with the preset grid neighborhood determination radius; If the Euclidean distance between the geometric centers is less than or equal to the grid neighborhood determination radius, then the two overlapping candidate points are determined to have spatial adjacency. Identify all overlapping candidate points that are interconnected in the spatial adjacency attribute and divide them into independent connected subsets; Count the number of overlapping candidate points in each connected subset, and retain connected subsets with a number greater than a preset effective area threshold as connected regions.

[0016] The beneficial effects of the technical solutions provided by the embodiments of the present invention include at least the following: By acquiring continuous image frames and extracting the light and shadow direction vector sequence of grid observation nodes, the light and shadow flow trend caused by micro-deformation of the skin surface is captured. The relative deviation between the main direction of static texture and the angle of dynamic light and shadow vector is used to characterize the real-time coupling degree between the adhesive structure and the skin texture. Stable continuous time segments are screened by differential variation amplitude to eliminate invalid tremor interference. The numerical variance is calculated based on the discreteness of the direction difference sequence to achieve a deep quantitative assessment of the dynamic stability of the target area. It provides accurate quantitative indicators reflecting the coordination of muscle and skin movement, and solves the problem that macro-contour monitoring is difficult to perceive the micro-mechanical interaction state. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a schematic diagram of the real-time monitoring system for the taping effect of thyroid cancer surgery neck and shoulder function rehabilitation provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the system framework of the present invention; Figure 3 This is a flowchart of the vector sequence generation module in this invention; Figure 4 This is a flowchart of the convergence section construction module in this invention; Figure 5 This is a flowchart of the direction difference calculation module in this invention; Figure 6 This is a flowchart of the stable section screening module in this invention; Figure 7 This is a flowchart of the effect monitoring module in this invention. Detailed Implementation

[0019] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0020] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0021] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0022] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0023] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0024] This invention provides a real-time monitoring system for the taping effect in the rehabilitation of neck and shoulder function after thyroid cancer surgery, such as... Figure 1 The diagram shown illustrates a real-time monitoring system for taping effects in neck and shoulder function rehabilitation after thyroid cancer surgery. The system includes: The vector sequence generation module acquires continuous image frames of the neck and shoulder after thyroid cancer surgery, divides the single frame image into a grid, extracts the image pixel points at the center of the grid as observation nodes, and determines the light and shadow direction vector sequence of adjacent observation nodes. The convergence segment construction module selects a specified observation node as the center position, calculates the angular deviation between the light and shadow direction vector sequences of adjacent observation nodes in the four-neighborhood of the specified observation node, and marks the light and shadow direction convergence segment. The orientation difference calculation module scans the brightness gradient within the grid of the observation node in the first frame of a series of images of the neck and shoulder after thyroid cancer surgery, compares it with the light and shadow orientation vector sequence, and filters the light and shadow-texture orientation difference sequence. The stable segment selection module determines the variation amplitude of the directional difference between adjacent time intervals in the light and shadow-texture directional difference sequence and selects stable segments of light and shadow-texture coupling. The effect monitoring module determines the coordinates of the target domain for pasting based on the spatial coordinate indices of the observation nodes marked as convergent segments in the light and shadow direction and stable segments in the light and shadow-texture coupling, quantifies the dynamic stability of the pasting area, and obtains the pasting effect monitoring results.

[0025] The light and shadow direction vector sequence includes time step number, node position identifier, and light and shadow vector angle value. The light and shadow direction convergence segment specifically includes convergence grid region index, four-neighbor angle deviation set, and convergence judgment threshold mark. The light and shadow-texture direction difference sequence includes texture gradient main direction value, texture flow field relative angle, and time series record list. The light and shadow-texture coupling stable segment includes first-order difference variation amplitude, relative static time window, and stable feature grid location. The bonding effect monitoring result specifically refers to the boundary coordinates of the aggregation area, sequence numerical variance term, and bonding effect monitoring index.

[0026] Please see Figure 2 and Figure 3 The vector sequence generation module includes: The node pixel extraction submodule acquires continuous image frames of the neck and shoulder after thyroid cancer surgery. It performs grid segmentation on each single frame of the continuous image frames of the neck and shoulder after thyroid cancer surgery, locates the center position of the grid and extracts the pixel point at the center position of the grid as a spatially fixed observation node, reads the gray value of each observation node on the single frame image, and generates the center pixel set of the observation node. The acquisition of continuous image frames of the neck and shoulder after thyroid cancer surgery was performed using a high frame rate industrial camera under controlled lighting conditions. The sampling frequency was set to 60 frames per second. This frequency setting was based on statistical analysis of the frequency data of physiological micro-tremors of the neck skin, calculating the maximum upper limit of involuntary muscle tremors on the skin surface. According to the Nyquist sampling theorem, the sampling frequency was set to more than twice this maximum frequency upper limit to prevent aliasing and ensure that details of skin surface micro-tremors and light and shadow changes with frequencies below 30 Hz could be captured. For each single frame of the continuous image frames of the neck and shoulder after thyroid cancer surgery, when performing grid segmentation, the grid size was set to a rectangular area of ​​10×10 pixels. This size setting was based on texture feature scale analysis of a large number of high-resolution neck skin texture images, measuring the average width of skin texture ridges and the groove spacing, calculating the arithmetic mean of these feature dimensions, and taking the smallest repeating unit size of the texture feature as the grid side length to ensure that each grid contains a single texture direction feature and avoid feature confusion caused by excessively large grids containing multiple texture directions. When locating the center of the grid, a geometric center calculation method is used to obtain the coordinates of the top-left and bottom-right vertices of the grid. Half the sum of the horizontal coordinates and half the sum of the vertical coordinates are then calculated to determine the coordinates of the intersection of the grid diagonals as a spatially fixed observation node. When reading the grayscale value of each observation node in a single frame image, the 8-bit depth grayscale value of the corresponding coordinate point in the image matrix is ​​directly accessed. This value quantifies the intensity of light reflection on the skin surface and ranges from 0 to 255. For example, in the 10th frame image, the observation node located at coordinates (100, 100) reads a grayscale value of 128. This value, along with its grayscale value of 125 in the 9th frame and 131 in the 11th frame, constitutes the temporal grayscale change sequence of this node. The set of grayscale values ​​of all observation nodes across all frames collectively generates the central pixel set of the observation node.

[0027] The brightness gradient comparison submodule extracts the brightness values ​​of the central pixel set of the observation node between temporally adjacent single-frame images, performs temporal difference processing, calculates the brightness change rate of the current observation node, performs gradient comparison with the brightness change rate of spatially adjacent observation nodes, determines the difference magnitude and positive / negative relationship between the two, and obtains the node brightness gradient comparison result. The brightness values ​​of the observed nodes in the central pixel set are extracted between temporally adjacent single-frame images. For example, the brightness value of the target observed node at the current moment is 150, and the brightness value at the previous moment was 148. During temporal difference processing, the brightness value at the current moment is subtracted from the brightness value at the previous moment; the calculation is 150 − 148 = 2, resulting in an instantaneous brightness change of 2. To calculate the brightness change rate of the current observed node, the instantaneous brightness change is divided by the time interval between adjacent frames, which is determined by the sampling frequency (e.g., 1 second divided by 60 frames is approximately 0.016 seconds). The calculation is 2 ÷ 0.016 = 125, resulting in a brightness change rate of 125 grayscale units per second for this node. Subsequently, the four spatially adjacent observed nodes (up, down, left, right) are selected, and the brightness change rates of these four adjacent nodes are obtained respectively. During gradient comparison, the brightness change rate of the current node is subtracted from the brightness change rates of its adjacent nodes. For example, if the current node's rate of change is 125 and the neighboring node's rate of change is 100, the difference between them is calculated as 125 - 100 = 25. After determining the difference and its sign, the brightness prominence state is judged based on the different ranges of the result value. Specifically, there are three possibilities: First, if the difference is positive (greater than 0), it indicates that the current node's brightness increases faster than its neighbors, and it is in a state of high-brightness accumulation. Second, if the difference is negative (less than 0), it indicates that the current node's brightness increases slower than its neighbors, and it is in a state of deepening shadows. Third, if the difference is equal to 0, it indicates that it is in a flat region where brightness changes synchronously. In this example, the calculated result 25 is greater than 0, so the node is determined to be in a state of high-brightness accumulation.

[0028] The vector sequence generation submodule determines the light and shadow vector angle representing the light and shadow flow trend on the skin surface at each observation node at the current moment based on the node brightness gradient comparison results. It then arranges the light and shadow vector angles in each consecutive frame image in chronological order to generate a light and shadow direction vector sequence. Based on the node brightness gradient comparison results, the light and shadow vector angle representing the light and shadow flow trend on the skin surface at each observation node at the current moment is determined. Specifically, the brightness gradient difference between horizontally adjacent nodes (left and right nodes) is used as the horizontal axis component (e.g., right gradient minus left gradient, value 10), and the brightness gradient difference between vertically adjacent nodes (upper and lower nodes) is used as the vertical axis component (e.g., upper gradient minus lower gradient, value 10). The arctangent function in the inverse trigonometric functions is used to calculate the vector angle synthesized from these two components. The calculation process is as follows: The angle value has several possibilities, and its quadrant is determined by the sign of its components: if both the horizontal and vertical components are positive, the angle is in the first quadrant (0 to 90 degrees); if the horizontal axis is negative and the vertical axis is positive, the angle is in the second quadrant (90 to 180 degrees); if both are negative, it is in the third quadrant (180 to 270 degrees); and if the horizontal axis is positive and the vertical axis is negative, it is in the fourth quadrant (270 to 360 degrees). In this example, both the horizontal and vertical axes are positive, so the angle is determined to be 45 degrees, located in the first quadrant. The light and shadow vector angles in each consecutive frame are arranged in chronological order, for example, the angle in the first frame is 45 degrees, the angle in the second frame is 46 degrees, the angle in the third frame is 47 degrees, and so on, until the last frame, thus generating a sequence of light and shadow direction vectors describing the trajectory of light and shadow direction changes over time at the observation node.

[0029] Please see Figure 2 and Figure 4 The convergence section construction module includes: The deviation calculation submodule selects a specified observation node as the center position and traverses the light and shadow direction vector sequence corresponding to the adjacent observation nodes in the four neighboring spatial domains of the specified observation node. For each time point in the sequence, it reads the light and shadow vector angle values ​​of the center position and the four neighboring positions respectively, calculates the angle deviation between the light and shadow vector angle of the observation node at the center position and the light and shadow vector angle of the observation nodes at the four neighboring positions, and generates a set of local direction angles. A specified observation node is selected as the center position, and its four neighboring observation nodes in the spatial quadrant are identified. The sequence of light and shadow direction vectors corresponding to these neighboring observation nodes is traversed. For each time point in the sequence, the light and shadow vector angle of the center observation node (e.g., 30 degrees) and the light and shadow vector angles of the four neighboring observation nodes (e.g., 32 degrees) are read. The angular deviation between the light and shadow vector angles of the center observation node and the four neighboring observation nodes is calculated; that is, the absolute value of the difference is calculated. The calculation process is as follows: Degrees. Perform this operation on all time points in the sequence. For example, if the sequence length is 1000 frames, calculate 1000 deviation values ​​and aggregate these deviation values ​​to generate a set of local orientation angles.

[0030] The threshold interval determination submodule compares each angle deviation within the local direction angle set with a preset convergence angle threshold, counts the frequency of each adjacent observation node's deviation value falling within the convergence angle threshold range throughout the entire time period of the sequence, and filters out the observation nodes whose corresponding frequency occupies the majority in the total time points to obtain the threshold node index. Each angular deviation within the local directional angle set is compared with a preset convergence angle threshold. This threshold is set based on micro-motion video data of healthy neck skin during natural breathing and slight swallowing movements. The data set of angular deviations between adjacent pixels is statistically analyzed, and the arithmetic mean and standard deviation of this set are calculated. The mean plus twice the standard deviation is taken as the threshold covering 95% of the normal fluctuation range, resulting in a set value of 15 degrees. Two possibilities exist during the comparison: first, the angular deviation is less than 15 degrees, which is considered a convergence moment, indicating high consistency of motion in the local area; second, the angular deviation is greater than or equal to 15 degrees, which is considered a divergence moment, indicating wrinkles or deformation in the local area. The frequency of each adjacent observation node's deviation falling within the convergence angle threshold range throughout the entire sequence is counted. For example, in 1000 total time points, if the deviation is less than 15 degrees at 850 time points, the frequency is 850. The selection process involves identifying observation nodes whose frequencies constitute a majority across all time points. This majority is determined by comparing the frequency percentage with a preset effective unidirectional ratio, set at 0.8. This ratio is derived from the statistical value of the minimum unidirectional ratio used to maintain structural continuity in skin biomechanics testing. The current percentage is calculated as 850 ÷ 1000 = 0.85. Comparing this result with the ratio threshold presents two possibilities: if the percentage is greater than 0.8, the node meets the selection criteria; if the percentage is less than or equal to 0.8, it does not. In this example, 0.85 > 0.8, therefore the node meets the selection criteria, resulting in the index of nodes meeting the threshold.

[0031] The convergence segment marking submodule locates the grid space region corresponding to the observation node that meets the threshold node index in the pixel grid coordinate system, identifies the adjacency relationship of the grid space region on the image plane, merges adjacent grid regions with the same flow direction in the spatial level, determines the geometric coverage of the merged region, and marks it as the attribute region with the same flow direction in the space, thus obtaining the light and shadow direction convergence segment. In the pixel grid coordinate system, locate the grid spatial region corresponding to the observation node that meets the threshold node index, and identify the adjacency relationship of the grid spatial region on the image plane. For example, if grid region A and grid region B are both determined to meet the threshold node, and the right boundary coordinates of grid region A coincide with the left boundary coordinates of grid region B, then they are considered to have an adjacency relationship. Spatially merge adjacent grid regions with consistent flow direction, that is, eliminate the geometric dividing line between grid region A and grid region B, and treat them as the same connected polygon region. Repeat this process until all adjacent and qualified grids are merged. Determine the geometric coverage of the merged region, for example, forming an irregular polygon covering the left neck region of the image. Mark the attribute region with consistent spatial flow direction, meaning that the skin in this region exhibits a consistent light and shadow flow characteristic during movement, and obtain the light and shadow direction convergence segment.

[0032] Please see Figure 2 and Figure 5 The direction difference calculation module includes: The texture direction extraction submodule selects the first frame image in the continuous image frames of the neck and shoulder after thyroid cancer surgery, performs brightness gradient scanning on all pixels in the grid area where the observation node is located in the first frame image, detects the gradient magnitude distribution in each direction in the area, and extracts the direction with the maximum gradient magnitude as a static index characterizing the inherent skin features of the area, which is then identified as the main direction of skin texture. The first frame of a series of images of the neck and shoulder after thyroid cancer surgery was selected. A brightness gradient scan was performed on all pixels within the grid area containing the observed node in the first frame. The grayscale gradients in the horizontal and vertical directions were calculated using an edge detection gradient operator. The gradient magnitude distribution in each direction within the detection area was analyzed. For example, the maximum gradient magnitude (5000) was detected at 30 degrees, while the minimum (500) was detected at 120 degrees. The direction of the maximum gradient magnitude (30 degrees) was extracted as a static indicator representing the inherent skin characteristics of the region. Based on the physiological characteristics of skin texture, the direction of the maximum gradient is usually perpendicular to the direction of the texture grooves. Therefore, the detected direction of the maximum gradient was rotated by 90 degrees to obtain the parallel direction of the texture. The calculation process is 30 + 90 = 120 degrees, which was confirmed as the main direction of the skin texture.

[0033] The relative angle calculation submodule takes the static skin texture main direction as the reference, traverses and reads the light and shadow vector angle data corresponding to each time point in the light and shadow direction vector sequence, calculates the relative angle value between the static texture main direction and the dynamic light and shadow vector angle, and generates a set of relative angle values. Using the static skin texture's principal direction as a reference, for example, a principal direction of 120 degrees. Iterate through and read the light and shadow vector angle data corresponding to each time point in the light and shadow direction vector sequence; for example, the light and shadow vector angle in frame 1 is 125 degrees. Calculate the relative angle between the static texture principal direction and the dynamic light and shadow vector angle, i.e., calculate the absolute value of the difference between the two, which is 120 - 125 = 5 degrees. Perform this calculation for all time points in the sequence, generating a set of relative angle values ​​containing the same number of angle differences as the number of frames.

[0034] The difference sequence filtering submodule arranges each relative angle value in the relative angle value set according to the time step order, records the change trajectory of the relative angle value in the time dimension, reflects the coupling relationship between the texture principal axis and the light and shadow flow direction of the skin surface, and generates a light and shadow-texture direction difference sequence. The process reflecting the coupling relationship between the principal axis of texture and the direction of light and shadow flow on the skin surface is as follows: Check the range of values ​​for each relative angle value in the set of relative angle values. If the relative angle value exceeds 90 degrees, perform angle complementation operation to convert it into an acute angle value within 90 degrees, and uniformly quantify the degree of deviation between the texture main axis and the light and shadow vector. Create a time series container indexed by time step, and store the converted acute angle values ​​into the time series container in the original time order in the light and shadow direction vector sequence. The generated time series container records the trajectory data of the deviation degree dynamically fluctuating over time. This numerical sequence represents the dynamic coupling state between the texture axis and the direction of light and shadow flow on the skin surface in the continuous time domain. Arrange each relative angle value in the relative angle value set according to the time step order. Check the value range of each relative angle value in the relative angle value set. There are two possibilities: First, if the relative angle value is greater than 90 degrees, perform angle complementation operation to convert it into an acute angle value within 90 degrees. For example, if the original calculated value of the relative angle is 100 degrees, it will be converted to 180 - 100 = 80 degrees. Second, if the relative angle value is less than or equal to 90 degrees, for example, if the original value is 80 degrees, it will remain unchanged. Create a time series container indexed by the time step, and store the converted acute angle values ​​into the time series container in the original time order in the light and shadow direction vector sequence. The generated time series container records the trajectory data of the deviation degree dynamically fluctuating over time. This numerical sequence represents the dynamic coupling state of the texture principal axis and the light and shadow flow direction on the skin surface in the continuous time domain.

[0035] Please see Figure 2 and Figure 6 The stable section filtering module includes: The variation amplitude calculation submodule performs first-order difference calculation on the direction difference values ​​between adjacent time points in the light and shadow-texture direction difference sequence, obtains the absolute value of the difference between the direction difference values ​​between adjacent time points, arranges them in the original time step order, and generates the variation amplitude of adjacent direction differences. The orientation difference values ​​between adjacent time points in the positioning lighting-texture orientation difference sequence are calculated using first-order difference. For example, the orientation difference value at time point T is 5 degrees, and the orientation difference value at time point T+1 is 8 degrees. The absolute value of the difference between the orientation difference values ​​between adjacent time points is obtained, and the calculation process is as follows: Degrees. Arranged in the original time step order, the variation amplitude of adjacent direction differences is generated. The result value of this step reflects the stability of the coupling relationship between light and shadow and texture, and is divided into different degrees of possibility: a larger value (e.g., greater than 5 degrees) indicates that the relative angle between the two changes drastically, corresponding to the skin surface being in a state of rapid deformation or trembling; a smaller value (e.g., less than 2 degrees) indicates that the relative angle between the two remains constant, corresponding to the skin surface being in a stable state.

[0036] The continuous segment recognition submodule introduces a preset stability judgment threshold as a baseline for measuring data fluctuations. It compares the variation amplitude of each adjacent direction difference with the stability judgment threshold, filters data points that are continuously less than the stability judgment threshold, extracts the continuous time intervals of the data points on the time axis, and confirms that the texture and the light and shadow flow on the skin surface remain relatively static within the time interval, thus obtaining stable continuous time segments. A preset stability threshold is introduced as a baseline for measuring data fluctuation. This threshold is set based on the noise data of light and shadow fluctuations of rigid objects (such as the clavicle without soft tissue coverage) in motion videos, calculating the maximum value of this noise amplitude, and setting it to 2 degrees. The amplitude of the difference between each adjacent direction is compared with the stability threshold, and data points that are consistently below the stability threshold are filtered out. For example, if the sequence contains consecutive values ​​of 1.5, 1.8, and 1.2, all less than 2 degrees. There are two possibilities for the comparison result: if the value is less than 2 degrees, it is determined to be a stable point; if the value is greater than or equal to 2 degrees, it is determined to be a fluctuating point. Continuous time intervals of data points are extracted on the time axis, for example, from frame 10 to frame 50. It is confirmed that the texture and the light and shadow flow on the skin surface remain relatively static within the time interval, resulting in a stable continuous time segment.

[0037] The stable segment marking submodule retrieves the observation nodes corresponding to stable continuous time segments, locates the geometric grid position of the target observation node in the spatial grid coordinate system, confirms the stability properties of multiple geometric grid positions in the process of light and shadow and texture interaction, and generates stable light and shadow-texture coupling segments. The process of confirming the stability properties of multiple geometric mesh positions during the interaction of light and shadow with texture is as follows: The total time span of all time segments determined to be stable and continuous at the same geometric grid location is calculated. Divide the total time span by the total time period length covered by the light and shadow-texture direction difference sequence to calculate the stability coverage ratio of the geometric mesh position in the time dimension; The stability coverage ratio is compared with the preset steady-state judgment benchmark ratio. When the stability coverage ratio is greater than or equal to the steady-state judgment benchmark ratio, it is confirmed that the geometric mesh position remains stable under the interaction of light and shadow flow and texture features; Retrieve the observation node corresponding to the stable continuous time segment and locate the geometric grid position of the target observation node in the spatial grid coordinate system. Calculate the total time span of all stable continuous time segments at the same geometric grid position. For example, if the node has three stable segments in the entire video, with lengths of 40, 60, and 100 frames respectively, the total is 40 + 60 + 100 = 200 frames. Divide the total time span by the total time period length covered by the light-shadow-texture direction difference sequence (e.g., 1000 frames) to calculate the stability coverage percentage of the geometric grid position in the time dimension. The calculation process is 200 ÷ 1000 = 0.2. Compare the stability coverage percentage with a preset steady-state judgment benchmark ratio, which is set to 0.6. This benchmark ratio is based on the statistical value of the minimum time coverage required for the adhesive surface to generate effective sustained stress in rehabilitation taping mechanical testing. The comparison result has two possibilities: first, if the ratio is greater than or equal to 0.6, the stable state is confirmed; second, if the ratio is less than 0.6, the unsteady state is confirmed. In this example, 0.2 < 0.6, so it belongs to the second possibility and this position is not marked; if the calculation result of another position is 0.7, then it belongs to the first possibility and generates a stable light and shadow-texture coupling section.

[0038] Please see Figure 2 and Figure 7 The effect monitoring module includes: The target domain coordinate determination submodule obtains the spatial coordinate indices of observation nodes marked as convergent segments of light and shadow direction and the spatial coordinate indices of observation nodes marked as stable segments of light and shadow-texture coupling. It performs an intersection filtering operation on the two sets of coordinate indices to determine overlapping candidate points. Based on the spatial adjacency distance of the overlapping candidate points in the grid array, it performs connectivity analysis and aggregates them to form connected regions, generating the target domain coordinates for patching. The process of performing connectivity analysis and aggregating overlapping candidate points into connected regions based on their spatial adjacency distance in the grid array is as follows: Calculate the Euclidean distance between the geometric centers of any two overlapping candidate points in the grid array; The Euclidean distance of the geometric center is numerically compared with the preset grid neighborhood determination radius; If the Euclidean distance between the geometric centers is less than or equal to the grid neighborhood determination radius, then the two overlapping candidate points are determined to have spatial adjacency. Identify all overlapping candidate points that are interconnected in the spatial adjacency attribute and divide them into independent connected subsets; Count the number of overlapping candidate points in each connected subset, and retain connected subsets with a number greater than a preset effective area threshold as connected regions; Obtain the spatial coordinate indices of observation nodes marked as convergent segments in the light and shadow direction and those marked as stable segments in the light and shadow-texture coupling. Perform an intersection filtering operation on the two sets of coordinate indices, retaining only coordinate points that exist in both sets simultaneously as overlapping candidate points. Perform connectivity analysis based on the spatial adjacency distance of overlapping candidate points in the grid array and aggregate them to form connected regions. Calculate the Euclidean distance between the geometric centers of any two overlapping candidate points in the grid array. Compare the Euclidean distance between the geometric centers with a preset grid neighborhood determination radius, which is set to 1.5 times the grid side length (e.g., 15 pixels), based on the statistical principle that allows for small gaps between grids but considers the whole grid as continuous. The comparison result has two possibilities: if the Euclidean distance between the geometric centers is less than or equal to 15 pixels, the two overlapping candidate points are determined to have spatial adjacency; if it is greater than 15 pixels, they are determined to be non-adjacent. Identify all interconnected overlapping candidate points in the spatial adjacency attribute and divide them into independent connected subsets. The number of overlapping candidate points in each connected subset is counted. Connected subsets with a number greater than a preset effective area threshold are retained as connected regions. This threshold is set to 5 nodes, based on the number of grid points corresponding to the minimum effective area of ​​mechanical effect produced by a single patch. If the number of subsets is greater than 5, they are retained; if it is less than or equal to 5, it is excluded as noise.

[0039] The stability quantization submodule iterates through the observation nodes that fall within the coordinate range of the target domain, extracts the light and shadow-texture direction difference sequence corresponding to each observation node and performs discreteness statistics, calculates the numerical variance of the sequence in the time dimension, quantifies the fluctuation of the light and shadow and texture coupling state in the tapping area, and obtains the dynamic stability value of the region. The observation nodes falling within the coordinate range of the target patching domain are traversed. The light and shadow-texture direction difference sequence corresponding to each observation node is extracted and its discreteness is statistically analyzed. The numerical variance of the sequence in the time dimension is calculated. First, the arithmetic mean of the sequence is calculated. For example, if the sequence values ​​are 6, 8, and 10, the calculation process is as follows: Then calculate the sum of squares of the differences between each value and the mean. The calculation process is as follows: Finally, divide the sum of squares by the number of data points. The calculation process is as follows: This value quantifies the fluctuation of the coupling state between light and shadow and texture within the patched area. The larger the value, the more intense the fluctuation, and the smaller the value, the smoother the fluctuation, thus obtaining the dynamic stability value of the area.

[0040] The monitoring results generation submodule introduces a preset rehabilitation training standard threshold as a benchmark parameter, compares the regional dynamic stability value with the rehabilitation training standard threshold, calculates the degree of deviation between the two values, assesses the recovery status of postoperative neck and shoulder function based on the degree of deviation, outputs quantitative indicators reflecting the coordination of neck and shoulder muscle and skin movement, and generates taping effect monitoring results. A preset rehabilitation training standard threshold is introduced as a benchmark parameter. This threshold is set based on the regional dynamic stability values ​​of the healthy side's neck and shoulder under the same movement. Multiple sets of healthy side data are averaged and set to 1.5. The regional dynamic stability value (e.g., 2.67) is compared with the rehabilitation training standard threshold (e.g., 1.5), and the degree of deviation between the two values ​​is calculated. The calculation process is as follows: The postoperative recovery status of neck and shoulder function is assessed based on the degree of deviation. The results are categorized into three possibilities: First, a deviation less than 0.2 indicates excellent recovery; second, a deviation between 0.2 and 0.5 indicates fair recovery; and third, a deviation greater than 0.5 indicates poor recovery or the presence of motor compensation. In this case, the calculated result of 0.78 is greater than 0.5, falling into the third possibility category, i.e., poor recovery. A quantitative index reflecting the coordination of neck and shoulder muscle and skin movement is output, generating a monitoring result for the taping effect.

[0041] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A real-time monitoring system for the effect of taping on neck and shoulder function rehabilitation after thyroid cancer surgery, characterized in that, The system includes: The vector sequence generation module acquires continuous image frames of the neck and shoulder after thyroid cancer surgery, divides the single frame image into a grid, extracts the image pixel points at the center of the grid as observation nodes, and determines the light and shadow direction vector sequence of adjacent observation nodes. The convergence segment construction module selects a specified observation node as the center position, calculates the angular deviation between the light and shadow direction vector sequences corresponding to the adjacent observation nodes in the four spatial neighborhoods of the specified observation node, and marks the light and shadow direction convergence segment. The orientation difference calculation module scans the brightness gradient within the grid of the first image observation node in a series of post-thyroid cancer surgery neck and shoulder images, compares it with the light and shadow orientation vector sequence, and filters the light and shadow-texture orientation difference sequence. The stable segment selection module determines the variation amplitude of the directional difference between adjacent time intervals in the light and shadow-texture directional difference sequence and selects stable light and shadow-texture coupling segments. The effect monitoring module determines the coordinates of the target domain for bonding based on the spatial coordinate indices of the observation nodes marked as the convergence zone of the light and shadow direction and the stable zone of the light and shadow-texture coupling, quantifies the dynamic stability of the bonding area, and obtains the bonding effect monitoring results.

2. The real-time monitoring system for taping effect in postoperative neck and shoulder function rehabilitation for thyroid cancer patients according to claim 1, characterized in that, The light and shadow direction vector sequence includes time step number, node position identifier, and light and shadow vector angle value. The light and shadow direction convergence segment specifically includes convergence grid region index, four-neighbor angle deviation set, and convergence judgment threshold mark. The light and shadow-texture direction difference sequence includes texture gradient principal direction value, texture flow field relative angle, and time series record list. The light and shadow-texture coupling stable segment includes first-order difference variation amplitude, relative static time window, and stable feature grid location. The bonding effect monitoring result specifically refers to the aggregation region boundary coordinates, sequence numerical variance term, and bonding effect monitoring index.

3. The real-time monitoring system for taping effect in thyroid cancer surgery neck and shoulder function rehabilitation according to claim 1, characterized in that, The vector sequence generation module includes: The node pixel extraction submodule acquires continuous image frames of the neck and shoulder after thyroid cancer surgery. It performs grid segmentation on each single frame of the continuous image frames of the neck and shoulder after thyroid cancer surgery, locates the center position of the grid and extracts the pixel point at the center position of the grid as a spatially fixed observation node, reads the gray value of each observation node on the single frame image, and generates the center pixel set of the observation node. The brightness gradient comparison submodule extracts the brightness values ​​of the observation nodes in the central pixel set of the observation node between temporally adjacent single-frame images, performs temporal difference processing, calculates the brightness change rate of the current observation node, performs gradient comparison with the brightness change rate of spatially adjacent observation nodes, determines the difference magnitude and positive / negative relationship between the two, and obtains the node brightness gradient comparison result. The vector sequence generation submodule determines the light and shadow vector angle representing the light and shadow flow trend on the skin surface at each observation node at the current moment based on the node brightness gradient comparison results, and arranges the light and shadow vector angles in each consecutive frame image in chronological order to generate a light and shadow direction vector sequence.

4. The real-time monitoring system for taping effect in thyroid cancer surgery neck and shoulder function rehabilitation according to claim 1, characterized in that, The convergence segment construction module includes: The deviation value calculation submodule selects a specified observation node as the center position and traverses the light and shadow direction vector sequence corresponding to the adjacent observation nodes in the four neighboring spatial domains of the specified observation node. For each time point in the sequence, it reads the light and shadow vector angle values ​​of the center position and the four neighboring positions respectively, calculates the angle deviation between the light and shadow vector angle of the observation node at the center position and the light and shadow vector angle of the observation nodes at the four neighboring positions, and generates a set of local direction angles. The threshold interval determination submodule compares each angle deviation in the set of local direction angles with a preset convergence angle threshold, counts the frequency of each adjacent observation node's deviation value falling within the convergence angle threshold range throughout the entire time period of the sequence, and filters the observation nodes whose corresponding frequency occupies the majority in the total time points to obtain the threshold node index. The convergence segment marking submodule locates the grid space region corresponding to the observation node that meets the threshold node index in the pixel grid coordinate system, identifies the adjacency relationship of the grid space region on the image plane, merges adjacent grid regions with the same flow direction in a spatial manner, determines the geometric coverage of the merged region, and marks it as an attribute region with the same flow direction in space to obtain the light and shadow direction convergence segment.

5. The real-time monitoring system for taping effect in postoperative neck and shoulder function rehabilitation for thyroid cancer patients according to claim 1, characterized in that, The orientation difference calculation module includes: The texture direction extraction submodule selects the first frame image in the continuous image frames of the neck and shoulder after thyroid cancer surgery, performs brightness gradient scanning on all pixels in the grid area where the observation node is located in the first frame image, detects the gradient magnitude distribution in each direction in the area, and extracts the direction with the maximum gradient magnitude as a static index characterizing the inherent skin features of the area, which is then identified as the main direction of skin texture. The relative angle calculation submodule takes the static main direction of the skin texture as a reference, traverses and reads the light and shadow vector angle data corresponding to each time point in the light and shadow direction vector sequence, calculates the relative angle value between the static main direction of the texture and the dynamic light and shadow vector angle, and generates a set of relative angle values. The difference sequence filtering submodule arranges each relative angle value in the relative angle value set according to the time step order, records the change trajectory of the relative angle value in the time dimension, reflects the coupling relationship between the texture main axis and the light and shadow flow direction of the skin surface, and generates a light and shadow-texture direction difference sequence.

6. The real-time monitoring system for taping effect in thyroid cancer surgery neck and shoulder function rehabilitation according to claim 5, characterized in that, The process reflecting the coupling relationship between the principal axis of texture and the direction of light and shadow flow on the skin surface is as follows: Check the range of values ​​for each relative angle value in the set of relative angle values. If the relative angle value exceeds 90 degrees, perform angle complementation operation to convert it into an acute angle value within 90 degrees, and uniformly quantify the degree of deviation between the texture main axis and the light and shadow vector. Create a time series container indexed by time step, and store the converted acute angle values ​​into the time series container in the original time order in the light and shadow direction vector sequence. The generated time series container records the trajectory data of the deviation degree dynamically fluctuating over time. This numerical sequence represents the dynamic coupling state between the texture axis and the direction of light and shadow flow on the skin surface in the continuous time domain.

7. The real-time monitoring system for taping effect in thyroid cancer surgery neck and shoulder function rehabilitation according to claim 1, characterized in that, The stable section filtering module includes: The variation amplitude calculation submodule locates the direction difference values ​​between adjacent time points in the light and shadow-texture direction difference sequence, performs first-order difference calculation, obtains the absolute value of the difference between the direction difference values ​​between adjacent time points, arranges them in the original time step order, and generates the variation amplitude of adjacent direction differences. The continuous segment recognition submodule introduces a preset stability judgment threshold as a baseline for measuring data fluctuations. It compares the variation amplitude of each adjacent direction difference with the stability judgment threshold, filters data points that are continuously less than the stability judgment threshold, extracts continuous time intervals of data points on the time axis, and confirms that the texture and the light and shadow flow on the skin surface remain relatively static within the time interval to obtain stable continuous time segments. The stable segment marking submodule retrieves the observation nodes corresponding to the stable continuous time segment, locates the geometric grid position of the target observation node in the spatial grid coordinate system, confirms the stability attributes of multiple geometric grid positions in the process of light and shadow and texture interaction, and generates a stable light and shadow-texture coupling segment.

8. The real-time monitoring system for taping effect in postoperative neck and shoulder function rehabilitation for thyroid cancer patients according to claim 7, characterized in that, The process of confirming the stability properties of multiple geometric mesh positions during the interaction of light and shadow with texture is as follows: The total time span of all time segments determined to be stable and continuous at the same geometric grid location is calculated. Divide the total time span by the total time period length covered by the light and shadow-texture direction difference sequence to calculate the stability coverage ratio of the geometric mesh position in the time dimension; The stability coverage ratio is compared with the preset steady-state judgment benchmark ratio. When the stability coverage ratio is greater than or equal to the steady-state determination benchmark ratio, it is confirmed that the geometric mesh position remains stable under the interaction of light and shadow flow and texture features.

9. The real-time monitoring system for taping effect in thyroid cancer surgery neck and shoulder function rehabilitation according to claim 1, characterized in that, The effect monitoring module includes: The target domain coordinate determination submodule obtains the spatial coordinate indexes of the observation nodes marked as the convergence section of the light and shadow direction and the spatial coordinate indexes of the observation nodes marked as the stable section of light and shadow-texture coupling. It performs an intersection filtering operation on the two sets of coordinate indices to determine overlapping candidate points. Based on the spatial adjacency distance of the overlapping candidate points in the grid array, it performs connectivity analysis and aggregates them to form a connected region, generating the target domain coordinates for bonding. The stability quantization submodule iterates through the observation nodes that fall within the coordinate range of the patching target domain, extracts the light-shadow-texture direction difference sequence corresponding to each observation node and performs discreteness statistics, calculates the numerical variance of the sequence in the time dimension, quantifies the fluctuation of the light-shadow and texture coupling state in the patching area, and obtains the dynamic stability value of the region. The monitoring results generation submodule introduces a preset rehabilitation training standard threshold as a benchmark parameter, compares the dynamic stability value of the region with the rehabilitation training standard threshold, calculates the degree of deviation between the two values, assesses the recovery status of postoperative neck and shoulder function based on the degree of deviation, outputs a quantitative index reflecting the coordination of neck and shoulder muscle and skin movement, and generates taping effect monitoring results.

10. The real-time monitoring system for taping effect in postoperative neck and shoulder function rehabilitation for thyroid cancer patients according to claim 9, characterized in that, The process of performing connectivity analysis and aggregating overlapping candidate points into connected regions based on their spatial adjacency distance in the grid array is as follows: Calculate the Euclidean distance between the geometric centers of any two overlapping candidate points in the grid array; The Euclidean distance of the geometric center is numerically compared with the preset grid neighborhood determination radius; If the Euclidean distance between the geometric centers is less than or equal to the grid neighborhood determination radius, then the two overlapping candidate points are determined to have spatial adjacency. Identify all overlapping candidate points that are interconnected in the spatial adjacency attribute and divide them into independent connected subsets; Count the number of overlapping candidate points in each connected subset, and retain connected subsets with a number greater than a preset effective area threshold as connected regions.

Images