Cervical spine care method and related apparatus based on state detection

Abstract

This application relates to the field of data processing, and provides a cervical spine care method and related device based on state detection. The method includes: acquiring personalized basic data of a target detection object, and acquiring cervical spine state-related data and cervical spine posture-related images of the target detection object; performing spatiotemporal feature fusion processing on the personalized basic data, cervical spine state-related data, and cervical spine posture-related images to obtain a spatiotemporal matrix of the target detection object's cervical spine state; performing cervical spine health assessment processing on the personalized basic data and the cervical spine state spatiotemporal matrix to obtain a cervical spine health status assessment result for the target detection object; generating cervical spine abnormality warning data for the target detection object when the non-healthy status assessment result in the cervical spine health status assessment result meets preset conditions; and formulating a target cervical spine care strategy based on the personalized basic data, cervical spine abnormality warning data, and non-healthy status assessment result to obtain a target cervical spine care strategy for the target detection object.

Description

Technical Field

[0001] This application relates to the field of data processing technology, specifically to a cervical spine care method and related device based on state detection. Background Technology

[0002] In today's society, the incidence of cervical spondylosis is on the rise and is increasingly affecting younger people. This is closely related to poor habits such as prolonged desk work and looking down at smart electronic devices. Currently, there are many detection and care solutions for cervical spondylosis, but all have certain shortcomings. Existing solutions typically rely on single-dimensional detection data recorded by a terminal to calculate the disease rate of the cervical spine and thus further determine its health. This approach not only suffers from significant time lag but also fails to provide real-time warnings of abnormal cervical postures, thereby reducing the effectiveness of the detection and care process for cervical spondylosis. Therefore, improving the effectiveness of the detection and care process for cervical spondylosis has become an urgent problem to be solved. Summary of the Invention

[0003] This application provides a cervical spine care method and related device based on state detection. It can combine the personalized basic data of the target detection object, as well as cervical spine state-related data and cervical spine posture-related images to construct a cervical spine state spatiotemporal matrix and formulate cervical spine care strategies based on the cervical spine state spatiotemporal matrix. This can result in more accurate, more appropriate and more personalized target cervical spine care strategies, thereby improving the effectiveness of the detection and care process for cervical spondylosis.

[0004] The first aspect of this application provides a cervical spine care method based on state detection, the method comprising:

[0005] Acquire personalized basic data of the target object, as well as cervical spine status-related data and cervical spine posture-related images of the target object collected by the sensor network;

[0006] Based on personalized basic data, cervical spine state-related data and cervical spine posture-related images, spatiotemporal feature fusion processing is performed to obtain the spatiotemporal matrix of the cervical spine state of the target detection object.

[0007] Based on personalized basic data and the spatiotemporal matrix of cervical spine status, cervical spine health assessment is performed to obtain the cervical spine health status assessment results of the target subjects.

[0008] When the non-health status assessment result in the cervical spine health status assessment meets the preset conditions, cervical spine abnormality warning data of the target detection object is generated.

[0009] Based on personalized basic data, cervical spine abnormality early warning data, and non-health status assessment results, a target cervical spine care strategy is developed to obtain the target cervical spine care strategy for the target detection subjects.

[0010] A second aspect of this application provides a cervical spine care device based on state detection, the device comprising:

[0011] The acquisition unit is used to acquire personalized basic data of the target detection object, as well as cervical spine state-related data and cervical spine posture-related images of the target detection object collected by the sensor network.

[0012] The first processing unit is used to perform spatiotemporal feature fusion processing based on the personalized basic data, the cervical spine state-related data and the cervical spine posture-related image to obtain the spatiotemporal matrix of the cervical spine state of the target detection object.

[0013] The second processing unit is used to perform cervical spine health assessment processing based on the personalized basic data and the cervical spine status spatiotemporal matrix to obtain the cervical spine health status assessment result of the target detection object.

[0014] The third processing unit is used to generate cervical spine abnormality warning data of the target detection object when the non-healthy status assessment result in the cervical spine health status assessment result meets the preset conditions.

[0015] The fourth processing unit is used to formulate a target cervical spine care strategy based on the personalized basic data, the cervical spine abnormality early warning data, and the unhealthy state assessment results, thereby obtaining the target cervical spine care strategy for the target detection object.

[0016] A third aspect of this application provides a terminal including a processor, an input device, an output device, and a memory, wherein the processor, input device, output device, and memory are interconnected, wherein the memory is used to store a computer program, the computer program including program instructions, and the processor is configured to invoke the program instructions to execute the step instructions as described in the first aspect of this application.

[0017] A fourth aspect of this application provides a computer-readable storage medium storing a computer program for electronic data interchange, wherein the computer program causes a computer to perform some or all of the steps described in the first aspect of this application.

[0018] A fifth aspect of this application provides a computer program product, wherein the computer program product includes a non-transitory computer-readable storage medium storing a computer program operable to cause a computer to perform some or all of the steps described in the first aspect of this application. The computer program product may be a software installation package.

[0019] Implementing the embodiments of this application has the following beneficial effects:

[0020] By acquiring personalized basic data of the target detection object, as well as cervical spine status-related data and cervical spine posture-related images collected by the sensor network, spatiotemporal feature fusion processing can be performed on the personalized basic data, the cervical spine status-related data, and the cervical spine posture-related images to obtain the spatiotemporal matrix of the target detection object's cervical spine status. Cervical spine health assessment processing can then be performed based on the personalized basic data and the spatiotemporal matrix of the cervical spine status to determine the cervical spine health status assessment result of the target detection object. When the non-healthy status assessment result in the cervical spine health status assessment result meets preset conditions, cervical spine abnormality early warning data for the target detection object can be generated. Furthermore, a target cervical spine care strategy can be formulated based on the cervical spine abnormality early warning data and the non-healthy status assessment result, thus obtaining the target cervical spine care strategy for the target detection object. This is beneficial to improving the effectiveness of the detection and care process for cervical spondylosis. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 This application provides a schematic diagram of the structure of a cervical spine care system based on state detection.

[0023] Figure 2 This application provides a schematic flowchart of a cervical spine care method based on state detection.

[0024] Figure 3 This is a schematic diagram of the structure of a terminal provided in an embodiment of this application;

[0025] Figure 4 This application provides a schematic diagram of the structure of a cervical spine care device based on state detection. Detailed Implementation

[0026] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0027] The terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or apparatuses.

[0028] In this application, the reference to "embodiment" means that a specific feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described in this application can be combined with other embodiments.

[0029] To better understand the cervical spine care method based on state detection provided in this application, a brief introduction to existing cervical spine care methods based on state detection is given below. In existing solutions, regarding data processing, some solutions rely solely on a single type of data, such as collecting cervical spine posture data only through an accelerometer. This fails to comprehensively reflect the health status of the cervical spine and lacks a comprehensive consideration of various aspects such as cervical spine structure and function. Furthermore, some solutions use relatively simple data processing methods, merely performing threshold comparisons and magnitude judgments, failing to fully extract effective information from the data, resulting in inaccurate results and an inability to accurately assess the complex state of the cervical spine. Regarding nursing strategy formulation, existing cervical spine care strategies often employ fixed patterns or preset treatment strategies, failing to fully consider individual differences among patients, such as cervical curvature, physical condition, and occupational characteristics, making it difficult to achieve truly personalized care. The determination of cervical spine abnormalities is not detailed or comprehensive enough, resulting in weak targeting of nursing strategies. For example, it may only focus on one aspect of cervical spine problems while ignoring other related abnormalities, failing to provide comprehensive care for cervical spine health as a whole. Therefore, the effectiveness of the detection and care process for cervical spondylosis is reduced.

[0030] To address the aforementioned issues, this application provides a cervical spine care method based on state detection. This method combines personalized basic data of the target detection object with cervical spine state-related data and cervical spine posture-related images to construct a cervical spine state spatiotemporal matrix. Based on this cervical spine state spatiotemporal matrix, a cervical spine care strategy can be formulated, resulting in a more accurate, appropriate, and personalized target cervical spine care strategy, thereby improving the effectiveness of the detection and care process for cervical spondylosis.

[0031] Please see Figure 1 , Figure 1 A schematic diagram of a cervical spine care system based on state detection is shown. Figure 1 As shown, a cervical spine care system based on state detection may include, but is not limited to, a data acquisition module, a spatiotemporal feature fusion module, a cervical spine health assessment module, a cervical spine abnormality early warning module, and a cervical spine care strategy formulation module. The data acquisition module collects personalized basic data, cervical spine state-related data, and posture images of the target object to provide comprehensive data support for subsequent analysis. The spatiotemporal feature fusion module extracts and fuses features from multiple sources in both time and space dimensions to form a spatiotemporal matrix reflecting the cervical spine state. The cervical spine health assessment module quantifies the cervical spine health state based on the spatiotemporal matrix through indicator comparison and risk grading. The cervical spine abnormality early warning module monitors the cervical spine state in real time, triggering early warnings and indicating the risk type and degree when data is abnormal. The cervical spine care strategy formulation module combines assessment and early warning results to match, optimize, and prioritize care strategies, generating personalized cervical spine care plans. These modules, following the logical route of "data acquisition → feature fusion → state assessment → risk early warning → strategy intervention," form a closed-loop process from passive monitoring to proactive intervention, significantly improving the intelligence, personalization, and effectiveness of cervical spine care.

[0032] Please see Figure 2 , Figure 2 This application provides a schematic flowchart of a cervical spine care method based on state detection. Figure 2 As shown, the state-detection-based cervical spine care method can be applied to a state-detection-based cervical spine care system. This method includes:

[0033] S10: Obtain personalized basic data of the target detection object, as well as cervical spine state-related data and cervical spine posture-related images of the target detection object collected by the sensor network.

[0034] The target detection subject can be used to indicate the individual who needs cervical spine care, that is, the individual who needs to adopt the state-based cervical spine care provided in this application. Personalized basic data can be used to indicate basic information related to the individual of the target detection subject. This personalized basic data may include, but is not limited to, basic physiological data, such as age, gender, height, weight, body mass index (BMI), past medical history (e.g., whether there is a previous cervical spine injury); basic behavioral data, such as occupation (programmer, teacher, athlete, etc.), daily habits (e.g., average daily sitting time, sleeping posture preference); and basic environmental data, such as work environment (whether long-term desk work is required), exercise frequency (number of times per week), etc., which can be used to analyze and evaluate the individual specificity of the target detection subject in subsequent steps.

[0035] A sensor network can be understood as a network of multiple sensors used to acquire data related to the cervical spine state and images related to the cervical spine posture of a target object. For example, a sensor network may include sensors that monitor cervical spine activity and muscle tension, as well as image acquisition devices that collect data on cervical spine posture and angles. Cervical spine state-related data can be used to indicate data related to the cervical spine state acquired through the sensor network, such as cervical spine motion data, force data, and physiological data. Cervical spine posture-related images can be used to indicate images of the cervical spine posture of the target object acquired through image acquisition devices in the sensor network, and can be used to analyze features such as cervical spine posture and angles.

[0036] S20: Based on the personalized basic data, the cervical spine state-related data, and the cervical spine posture-related images, spatiotemporal feature fusion processing is performed to obtain the spatiotemporal matrix of the cervical spine state of the target detection object.

[0037] The cervical spine state spatiotemporal matrix can be understood as a data structure that integrates personalized basic data, temporal and spatial features of cervical spine state-related data, and cervical spine posture-related image features. It can be used to more comprehensively describe the cervical spine state of the target object.

[0038] Specifically, by acquiring personalized basic data of the target detection object, cervical spine state-related data, and cervical spine posture-related images, weights are first calculated on the personalized basic data, temporal and spatial features are extracted from the cervical spine state data, and posture angle features are extracted from the posture images. These are then fused to form a spatiotemporal matrix of cervical spine state. Detailed descriptions can be found in the following embodiments, and will not be repeated here.

[0039] By fusing and extracting features from multiple data sources, a multidimensional correlation can be achieved between physiological characteristics (such as curvature), behavioral data (such as prolonged sitting), and individual differences (such as occupation). This allows the spatiotemporal matrix of cervical spine status to capture fatigue accumulation trends over time, locate abnormal posture areas in space, and highlight key influencing factors through personalized weights (e.g., higher weighting for prolonged sitting in programmers). This provides comprehensive and individualized feature support for subsequent health assessments, improving the accuracy and relevance of cervical spine status analysis.

[0040] S30: Based on the personalized basic data and the cervical spine status spatiotemporal matrix, perform cervical spine health assessment processing to obtain the cervical spine health status assessment result of the target detection object.

[0041] Cervical spine health assessment can be understood as a process of analyzing and judging the cervical spine health status of a target subject based on personalized baseline data and a cervical spine state spatiotemporal matrix. The cervical spine health status assessment results can be used to indicate the cervical spine health status of the target subject after the cervical spine health assessment is performed based on the personalized baseline data and the cervical spine state spatiotemporal matrix.

[0042] Optionally, the cervical spine health status assessment results may include, but are not limited to, time risk assessment results, spatial risk assessment results, and personalized risk assessment results. Among them, time risk assessment results may include, for example, a sedentary risk index, which is calculated as: average sedentary time (minutes) × 0.01 × occupational sedentary coefficient (where 1.5 can be used for programmers and 1.2 for teachers); spatial risk assessment results may include, for example, a curvature abnormality index, which is calculated as: (Deflection of curvature of each segment × segment weight), where the weight of C5-C6 is 0.4 (more prone to wear and tear), etc., this application does not impose any restrictions on this.

[0043] Optionally, the cervical spine health status assessment results may also include a cervical spine health status score. Specifically, the cervical spine health status score = 100. (Z1 temporal risk + Z2 spatial risk + Z3 personalized risk), where Z1, Z2, and Z3 can represent the weights of temporal risk, spatial risk, and personalized risk, respectively, and this application does not impose any restrictions on this. Optionally, the grading criteria for cervical spine health status can be: cervical spine health status score ≥ 90, indicating a healthy cervical spine; 80 ≤ cervical spine health status score < 90, indicating a mildly abnormal cervical spine; 70 ≤ cervical spine health status score < 80, indicating a moderately abnormal cervical spine; cervical spine health status score < 70, indicating a severely abnormal cervical spine, and this application does not impose any restrictions on this.

[0044] For example, taking a 30-year-old programmer A as an example, if programmer A's average sedentary time is 60 minutes, the sedentary coefficient for the programmer profession is 1.5, the deviation of the spinal curvature of the C5-C6 segment is +12°, and the average deviation of other segments is 0°, then programmer A's time risk can be: 60 minutes × 0.01 (unit standardized factor) × 1.5 = 0.9, and the spatial risk can be: 12° × 0.4 + 0° × 0.6 = 4.8; if the time risk weight is 0.35, the spatial risk weight is 0.37, the personalized risk weight is 0.28, and the personalized risk is 0.3 (which can be obtained by quantifying the age of 30), then programmer A's health score = 100 - (0.35 × 0.9 + 0.37 × 4.8 + 0.28 × 0.3) ≈ 97.3 points, and this application does not impose any restrictions on this.

[0045] S40: When the non-healthy status assessment result in the cervical spine health status assessment result meets the preset conditions, generate cervical spine abnormality early warning data for the target detection object.

[0046] The unhealthy state assessment result can be understood as the unhealthy assessment result indicated in the cervical spine health status assessment result, such as the "prolonged sitting" assessment result indicated in the cervical spine health status assessment result. Cervical spine abnormality early warning data can be used to indicate that when the unhealthy state assessment result in the cervical spine health status assessment result meets preset conditions, such as the duration of "prolonged sitting" exceeding 2 hours, data information is generated to indicate that there is an abnormality in the cervical spine status of the target detection object.

[0047] It should be noted that the preset conditions can be understood as pre-set rules for abnormal judgment, which can be used to determine whether the cervical spine health assessment results trigger an early warning. Optionally, the preset conditions can be a single-dimensional early warning threshold, such as continuous sitting for ≥90 minutes, which is not limited in this application. Optionally, the preset conditions can be multi-dimensional early warning thresholds, such as multiple time-dimensional early warning thresholds, such as continuous sitting for ≥90 minutes and daily exercise frequency = 0; multiple spatial-dimensional early warning thresholds, such as the curvature deviation of the C5-C6 segment of the spine ≥15° and head forward tilt ≥30°; or multiple comprehensive-dimensional early warning thresholds, such as a health score <80 points and a personalized risk coefficient ≥0.5, or continuous sitting for ≥90 minutes and the curvature deviation of the C5-C6 segment of the spine ≥15°, etc., which is not limited in this application.

[0048] For example, if the non-health status assessment results of programmer A's cervical spine health status assessment are as follows: continuous sitting for 120 minutes (time dimension warning triggered), the curvature of the C5-C6 segment of the spine deviates by 18° (spatial dimension warning triggered), and the health score is 75 points (comprehensive score dimension warning triggered), then cervical spine abnormality warning data can be generated for programmer A, such as "continuous sitting exceeds the threshold, abnormal spinal curvature, it is recommended to adjust immediately," etc. This application does not impose any restrictions on this.

[0049] Optionally, cervical spine abnormality warning data can be classified into levels, such as establishing a three-level warning system. For example, different warning colors (the color of warning signal lights or the color of warning text) can be used to classify cervical spine abnormality warning data into levels. For instance, yellow warning data can be used to indicate mild cervical spine abnormality, such as mild abnormality in a single dimension (e.g., sitting for 90 minutes); orange warning data can be used to indicate moderate abnormality, such as mild abnormality in multiple dimensions or moderate abnormality in a single dimension (e.g., sitting for 120 minutes + head tilting forward 25°); and red warning data can be used to indicate severe abnormality, such as moderate abnormality in multiple dimensions or severe abnormality in a single dimension (e.g., health score <70 or curvature deviation ≥20°). This application does not impose any restrictions on this.

[0050] S50: Based on the cervical spine abnormality early warning data and the non-health status assessment results, a target cervical spine care strategy is formulated to obtain the target cervical spine care strategy for the target detection object.

[0051] A targeted cervical spine care strategy can be understood as a specific care plan developed for the target individual based on cervical spine abnormality early warning data and cervical spine abnormality status assessment results. Optionally, this targeted cervical spine care strategy may include, but is not limited to, nursing measures and rehabilitation suggestions.

[0052] Specifically, in the process of formulating a target cervical spine care strategy, the type of cervical spine abnormality can be determined by cervical spine abnormality early warning data and non-health status assessment results. Then, a corresponding care strategy can be matched from the target cervical spine care strategy library. This allows for the optimization of strategy parameters (such as adjusting traction time based on curvature deviation) based on the cervical spine abnormality early warning data and non-health status assessment results. Furthermore, the strategy priority can be ranked by combining the personalized basic data of the target detection subject (such as occupation and age) to generate the target cervical spine care strategy. It should be noted that detailed descriptions can be found in the following embodiments, and will not be repeated here.

[0053] By integrating the personalized basic data of the target subjects (such as age, occupation, and body type mentioned above) with cervical spine status-related data and cervical spine posture-related images collected by sensors, a spatiotemporal matrix of cervical spine status that incorporates spatiotemporal features can be further constructed, realizing a closed-loop process from "individual differences" to "status assessment." Specifically, multi-source data fusion can improve the comprehensiveness of cervical spine health assessment, spatiotemporal matrix-based analysis can capture dynamic changes and spatial abnormalities in cervical spine status, and combined with pre-set abnormality warnings and customized nursing strategies, an automated health management system of "detection-assessment-intervention" can be formed, which can effectively improve the accuracy and timeliness of cervical spine care for the target subjects.

[0054] In this embodiment, by acquiring personalized basic data of the target detection object, and acquiring cervical spine status-related data and cervical spine posture-related images of the target detection object collected by the sensor network, spatiotemporal feature fusion processing can be performed based on the personalized basic data, the cervical spine status-related data, and the cervical spine posture-related images to obtain the spatiotemporal matrix of the target detection object's cervical spine status. Cervical spine health assessment processing can then be performed based on the personalized basic data and the spatiotemporal matrix of the cervical spine status to determine the cervical spine health status assessment result of the target detection object. Therefore, when the non-healthy status assessment result in the cervical spine health status assessment result meets preset conditions, cervical spine abnormality warning data for the target detection object can be generated. Furthermore, a target cervical spine care strategy can be formulated based on the cervical spine abnormality warning data and the non-healthy status assessment result, thus obtaining the target cervical spine care strategy for the target detection object, which is beneficial to improving the effectiveness of the detection and care process for cervical spondylosis.

[0055] In one possible implementation, the spatiotemporal feature fusion processing can be achieved through a spatiotemporal feature fusion module in a state-detection-based cervical spine care system. Specifically, firstly, personalized weights are calculated on the personalized basic data, and state temporal and spatial features are extracted from the cervical spine state-related data, as well as posture angle features are extracted from the cervical spine posture-related images. Then, the calculated and extracted feature data are subjected to spatial feature fusion and spatiotemporal fusion processing to finally obtain the spatiotemporal matrix of the cervical spine state of the target detection object. Specifically, a method for obtaining the spatiotemporal matrix of the cervical spine state of the target detection object by performing spatiotemporal feature fusion processing based on the personalized basic data, the cervical spine state-related data, and the cervical spine posture-related images includes:

[0056] A1. Perform personalized weight calculation on the personalized basic data to obtain a personalized weight vector;

[0057] A2. Extract the temporal features of the cervical spine state-related data to obtain the cervical spine state temporal feature vector;

[0058] A3. Extract state space features from the cervical spine state-related data to obtain the cervical spine state space feature vector;

[0059] A4. Extract posture angle features from the cervical spine posture-related images to obtain cervical spine posture angle feature vectors;

[0060] A5. Perform spatial feature fusion processing on the cervical spine state spatial feature vector and the cervical spine posture angle feature vector to obtain the cervical spine fused spatial feature vector.

[0061] A6. Perform spatiotemporal fusion processing on the personalized weight vector, the cervical spine state temporal feature vector, and the cervical spine fusion spatial feature vector to obtain the spatiotemporal matrix of the cervical spine state of the target detection object.

[0062] Personalized weight calculation can be understood as the process of assigning corresponding weights to personalized basic data based on its characteristics and importance. Specifically, the personalized basic data can first be quantized, and then the quantized feature vector can be linearly transformed. Next, feature mapping can be performed using a feature importance mapping matrix, thereby normalizing the resulting personalized mapped feature vector. Finally, the result of personalized weight calculation, i.e., the personalized weight vector, can be obtained.

[0063] This personalized weight vector can be understood as a set of numerical vectors generated based on individual differences in characteristics. It can be used to quantify the importance of different characteristics, enabling personalized data analysis and evaluation in subsequent steps. For example, taking personalized base data as [age, occupation, body type], the personalized weight vector can be calculated using personalized weight calculation processing. ,in, It can be a personalized weight vector corresponding to age. It can be a personalized weight vector corresponding to a profession. This application does not impose any restrictions on the personalized weight vector corresponding to body type.

[0064] The cervical spine state temporal feature vector can be understood as a vector extracted from cervical spine state-related data to reflect time-series characteristics. This vector can contain relevant feature information about the changes in cervical spine state over time. Specifically, firstly, multi-dimensional temporal features can be extracted from the cervical spine state-related data. Then, the extracted time-series feature vectors are concatenated, and further, long-term and short-term feature abstraction iterative processing is used to abstract and iterate the concatenated feature vector. Thus, the final target abstract memory feature vector can be used as the cervical spine state temporal feature vector. This application does not impose any limitations on this process.

[0065] The cervical spine state space feature vector can be understood as a vector extracted from cervical spine state-related data to describe the spatial state characteristics of the cervical spine. The cervical spine state space feature vector can include spatial features such as spinal curvature.

[0066] Optionally, state-space feature extraction of the cervical spine state-related data may include, but is not limited to, calculation of spinal curvature and muscle tension; this application does not impose any limitations on this. Specifically, for the calculation of spinal curvature, the spinal curvature data of segments C3-C7 of the target object can be obtained through a posture sensor. For example, if spinal curvature data is obtained... ,in, The actual spinal curvature value of the third cervical segment. The actual spinal curvature value of the fourth cervical segment. The actual spinal curvature value of the 5th cervical segment. The actual spinal curvature value of the 6th cervical segment. The actual spinal curvature value of the 7th cervical segment; and the difference from the normal spinal curvature is calculated sequentially. ,in, It can represent the absolute difference between the actual curvature and the normal curvature of the i-th cervical vertebra. It can represent the actual curvature of the i-th cervical vertebra. It can represent normal curvature. Muscle tension can be calculated by calculating the root mean square (RMS) value of the electromyography (EMG) signal. ,in, The value of the electromyography signal at the i-th sampling point can be represented by n, which can represent the total number of sampling points of this electromyography signal. The root mean square value of the electromyography signal is used as the feature vector of muscle tension. This application does not impose any restrictions on this.

[0067] The cervical spine posture angle feature vector can be understood as a feature vector representing cervical spine posture angle information obtained through feature analysis and extraction of cervical spine posture-related images. Specifically, an image acquisition device can be used to acquire a back image of the target object (other images are also acceptable, this application does not limit this), and a residual network can be used to extract spatial features from the back image. Optionally, the coordinates of key points in the back image can be determined, such as the pixel coordinates of key positions like the cervical vertebrae, scapula, and lumbar vertebrae. ,in, This can be the coordinates of the first key point (such as the apex of the cervical spine). This could be the coordinates of a second key point (such as the scapula), etc. The coordinates can be the coordinates of the k-th keypoint, where k can be the total number of keypoints on the body to be detected; this application does not impose any limitation on this. Furthermore, posture angles can be calculated based on the keypoint coordinates, for example, calculating the head tilt angle. Shoulder tilt However, this application does not impose any limitations on this. For example, the forward tilt angle can be calculated using geometric transformation formulas. ,in, It can be the coordinates of the head's center of gravity. It can be the coordinates of the neutral position of the shoulder.

[0068] The spatial feature fusion processing of the cervical spine state space feature vector and the cervical spine posture angle feature vector can be understood as concatenating the cervical spine state space feature vector and the cervical spine posture angle feature vector. For example, concatenating the cervical spine state space feature vector and the cervical spine posture angle feature vector according to their dimensions can yield the cervical spine fusion spatial feature vector.

[0069] The cervical spine fusion spatial feature vector can be understood as the feature vector obtained by fusing the cervical spine state spatial feature vector and the cervical spine posture angle feature vector. In other words, by fusing the cervical spine state spatial feature vector and the cervical spine posture angle feature vector, we can combine the feature information of the two feature vectors in the spatial dimension and obtain more comprehensive and accurate feature information in the spatial dimension.

[0070] For example, the cervical spine posture angle feature vector includes two dimensions: key point coordinates and posture angle. After extracting the key point coordinates and posture angle, the cervical spine posture angle feature vector can be further obtained. For example, the spatial position features of key points such as the scapula and cervical vertebrae in the back image can be obtained.

[0071] Optionally, the process of fusing the cervical spine state spatial feature vector and the cervical spine posture angle feature vector to obtain the cervical spine fused spatial feature vector can be seen in the following formula:

[0072]

[0073] in, It can be used to represent the spatial feature vector of cervical spine fusion; It can be used to represent the feature vector of cervical spine posture angle; It can be used to represent the spinal curvature feature vector in the cervical spine state space feature vector; It can be used to represent the operation of spatial feature fusion processing, and in this embodiment, it can represent the operation of splicing. It can be used to represent the first dimension of the cervical spine posture angle feature vector; It can be used to represent the second dimension of the cervical spine posture angle feature vector; It can be used to represent the 512th dimension of the cervical spine posture angle feature vector; 512 can be used to represent the total number of dimensions of the cervical spine posture angle feature vector in this embodiment. It can be used to represent the spinal curvature feature vector of the third cervical vertebra in the cervical state space feature vector; It can be used to represent the spinal curvature feature vector of the fourth cervical vertebra in the cervical state space feature vector; It can be used to represent the spinal curvature feature vector of the 5th cervical vertebra in the cervical spine state space feature vector; It can be used to represent the spinal curvature feature vector of the 6th cervical vertebra in the cervical state space feature vector; It can be used to represent the spinal curvature feature vector of the 7th cervical vertebra in the cervical state space feature vector.

[0074] It should be noted that the above formula uses the example of the cervical spine state space feature vector including the spinal curvature feature vector, and does not constitute a limitation on this application. Optionally, the cervical spine state space feature vector may also include muscle tension feature vectors, etc., and this application does not impose any restrictions on this. For the relevant content of spatial feature fusion processing of muscle tension feature vectors, please refer to the detailed description of spatial feature fusion processing of spinal curvature feature vectors, which will not be repeated here.

[0075] Spatiotemporal fusion processing can be understood as the operation of integrating and fusing data from different sources and of different types in the time and space dimensions to obtain a more comprehensive, accurate and representative spatiotemporal matrix of cervical spine status.

[0076] Optionally, the process of performing spatiotemporal fusion processing on the personalized weight vector, the cervical spine state temporal feature vector, and the cervical spine fusion spatial feature vector to obtain the spatiotemporal matrix of the cervical spine state of the target detection object can be found in the following formula:

[0077]

[0078] in, W can represent the spatiotemporal matrix of the cervical spine state of the target object; W can represent the personalized weight vector. This can represent a weighted dot product operation; It can be used to represent operations that perform spatiotemporal fusion processing; in this embodiment, it can represent operations that perform vector concatenation. It can represent the temporal feature vector of cervical spine status; It can represent the spatial feature vector of cervical spine fusion.

[0079] In the spatiotemporal feature fusion process provided in this application embodiment, in-depth data mining and structured representation can be achieved through personalized weight calculation, temporal and spatial feature extraction, and multi-dimensional spatiotemporal fusion processing. Specifically, the personalized weight vector can dynamically adjust the importance of indicators such as prolonged sitting time and cervical curvature based on individual characteristics (e.g., programmers have a higher weight for prolonged sitting); temporal feature extraction can capture long-term trend features such as fatigue accumulation; and spatial feature fusion can integrate the spatial correlation between spinal curvature and posture angles. The resulting spatiotemporal matrix of cervical spine status reflects the temporal evolution of cervical spine status and can locate spatial abnormal areas, thus providing multi-dimensional feature support for accurate assessment.

[0080] In one possible implementation, personalized weight calculation can be performed using a personalized weight calculation module within the spatiotemporal feature fusion module. Specifically, a personalized weight vector can be generated using a feature importance mapping matrix and personalized basic data. For example, the personalized basic data can first be processed into feature vectors that are easy to process and analyze, then linearly transformed, followed by feature mapping using the feature importance mapping matrix, and finally normalized to obtain the personalized weight vector. Specifically, a method for performing personalized weight calculation on the personalized basic data to obtain a personalized weight vector includes:

[0081] B1. Quantize the personalized basic data to obtain a personalized basic feature vector;

[0082] B2. Perform a linear transformation on the personalized basic feature vector to obtain a personalized linear feature vector;

[0083] B3. Through the feature importance mapping matrix, the personalized linear feature vector is subjected to feature mapping processing to obtain the personalized mapped feature vector;

[0084] B4. Normalize the personalized mapping feature vector to obtain the personalized weight vector.

[0085] Quantization can be understood as transforming personalized basic data into quantifiable numerical or vector forms to facilitate computer processing and analysis. Personalized basic feature vectors can be used to indicate the feature vectors that represent the personalized basic data after quantization.

[0086] For example, let's take the quantification of personalized basic data such as age, occupation, and body type as an example. For age, you can refer to the formula. The age is quantified, among which, Age can represent the normalized age feature value, while Age can represent the actual age of the target object being detected. This can represent the minimum threshold value for age quantification. This can represent the maximum threshold for age quantification. For occupational data, an occupational quantification vector can be generated using one-hot encoding. Where Occ can represent the occupation quantization vector obtained after one-hot encoding of the occupation. This can represent the feature value in the occupational quantification vector that represents the first type of occupation (such as office worker). This can represent the feature value in the occupational quantification vector that represents the second type of occupation (such as an athlete). The BMI value can represent the feature value of the nth occupation category in the occupational quantification vector, where n can represent the occupation type; for body type, the BMI value can be expressed using the formula... This is to quantify the BMI value, where, BMI can be represented by the normalized characteristic value after standardization. BMI can represent the actual body mass index of the target subject. It can represent the average BMI of a preset reference population. It can represent the standard deviation of BMI for a pre-defined reference population, but this application does not impose any restrictions on it.

[0087] Linear transformation processing can be understood as performing linear operations on personalized basic feature vectors to change their dimensions or feature representation, thereby better adapting to the needs of subsequent calculations and analysis. Optionally, this application embodiment uses the transpose operation of personalized basic feature vectors as an example for illustration, to better meet the dimensional requirements of feature mapping processing in subsequent steps, and does not constitute a limitation on this application.

[0088] A feature importance mapping matrix can be used to measure the importance of each feature in a personalized basic feature vector and map it to a new feature space. In other words, the feature importance mapping matrix reflects the influence coefficient of the personalized basic feature vector on each feature type (i.e., features in the time dimension or features in the spatial dimension). This feature importance mapping matrix can be trained from sample personalized basic data, and this application does not impose any restrictions on this. Optionally, the feature importance mapping matrix can also be understood as a feature importance mapping function to map personalized basic features to a feature importance space, thereby achieving a semantic transformation from personalized basic features to importance weights, and this application does not impose any restrictions on this either.

[0089] Optionally, the feature mapping processing in this embodiment can also be understood as performing matrix multiplication between the feature importance mapping matrix and the personalized linear feature vector. This achieves a weighted summation of the personalized basic feature vector and the feature importance mapping matrix, thereby realizing a linear transformation from basic features to feature importance. The result is a linear combination vector of feature importance, i.e., the personalized mapping feature vector. This personalized mapping feature vector can be used to indicate the feature vector that highlights the differences in importance between different features after processing the personalized linear feature vector through the feature importance mapping matrix.

[0090] Normalization can be understood as adjusting the numerical range of personalized mapping feature vectors to meet certain numerical range and distribution requirements. Optionally, normalization can typically transform personalized mapping feature vectors into probability distributions or unit vectors for comparison and calculation in subsequent steps.

[0091] Optionally, the process of calculating personalized weights on the personalized basic data to obtain a personalized weight vector can be found in the following formula:

[0092]

[0093] Where W can represent the personalized weight vector; softmax() represents the normalization operation; X can represent a feature importance mapping matrix; X can represent a personalized basic feature vector. b can represent the linear transformation result of the personalized basic feature vector, and in this application, it can be the transpose of the personalized basic feature vector; b can represent the bias term, which can be used as the benchmark for adjusting the weights.

[0094] In this embodiment, through a weight calculation process involving quantization, linear transformation, feature mapping, and normalization, individual personalized basic data is transformed into a computable weight vector, achieving digital mapping of personalized basic data. In other words, through a feature importance mapping matrix, experiential knowledge such as "the influence of age on cervical curvature" and "the correlation between occupation and sedentary behavior" is encoded into matrix parameters. After linear transformation and normalization, the weight vector conforms to a probability distribution (and the weight sum is 1), thus ensuring the comparability of feature importance across different individuals. This personalized weight calculation module allows the spatiotemporal feature fusion model to adapt to the physiological characteristics and behavioral habits of different users, avoiding a "one-size-fits-all" assessment and improving the scientific rigor of personalized care.

[0095] In one possible implementation, the state-time feature extraction can be achieved through the state-time feature extraction module within the spatiotemporal feature fusion module. Specifically, firstly, multi-dimensional time-time features can be extracted from cervical spine state-related data to obtain a set of multiple time-time feature vectors of different dimensions. Then, these multiple time-time feature vectors are concatenated to obtain a concatenated time-time feature vector. Next, long-short-term feature abstraction iterative processing is used to perform feature abstraction iterative processing. Finally, the feature vector obtained from the final iterative processing is used as the cervical spine state time-time feature vector. Specifically, a method for extracting state-time features from the cervical spine state-related data to obtain a cervical spine state time-time feature vector includes:

[0096] C1. Perform multi-dimensional temporal feature extraction processing on the cervical spine state-related data to obtain a reference temporal feature vector set;

[0097] C2. Concatenate each reference time-series feature vector in the reference time-series feature vector set to obtain the cervical spine concatenated time-series feature vector;

[0098] C3. Perform long-short-term feature abstraction and iterative processing on the cervical spine splicing temporal feature vector to obtain the target abstract memory feature vector;

[0099] C4. Use the target abstract memory feature vector as the cervical spine state temporal feature vector.

[0100] The multi-dimensional temporal feature extraction process can be understood as extracting features from cervical spine state-related data over time from multiple aspects. For example, the mean, standard deviation, and peak value of the prolonged sitting time of the target object can be extracted from the cervical spine state-related data; and feature vectors of multiple dimensions such as the movement frequency and continuous head-down time of the target object can be extracted from the cervical spine state-related data. This application does not impose any limitations on this. The reference temporal feature vector set can be understood as a set composed of multiple temporal feature vectors of different dimensions extracted from cervical spine state-related data.

[0101] For example, by extracting sedentary duration features from cervical spine status-related data, a sedentary duration sequence can be obtained. (Unit: minutes), of which, This can represent the duration of sitting recorded at the first time point in the sedentary duration sequence. This can represent the duration of sitting recorded at the second time point in the sedentary duration sequence. This can represent the sitting duration recorded at the m-th time point in the sedentary duration sequence, where m represents the total number of time points in the sedentary duration sequence; further, the mean of the sedentary duration can be extracted. Standard deviation of sedentary time Peak of continuous sedentary behavior Optionally, motion frequency features can be extracted from cervical spine status-related data to obtain motion frequency. It can be further standardized to (i.e., the feature values ​​after standardization of movement frequency); by extracting continuous head-down time features from cervical spine state-related data, a continuous head-down duration sequence can be obtained. (Unit: minutes), of which, It can represent the duration of consecutive head-down movements recorded at the first time point in a sequence of consecutive head-down movements. It can represent the duration of continuous head-down movement recorded at the second time point in a sequence of continuous head-down movement durations. It can represent the duration of continuous head-down movement recorded at the first time node in a continuous head-down duration sequence, and p can represent the total number of time nodes contained in the continuous head-down duration sequence.

[0102] The cervical spine splicing temporal feature vector can be used to indicate the feature vector formed by splicing together the various reference temporal feature vectors in a certain order (such as chronological order). It can be understood that this cervical spine splicing temporal feature vector integrates temporal feature information from multiple dimensions.

[0103] For example, the above reference temporal features can be concatenated in chronological order to form a cervical spine splicing temporal feature vector of uniform dimension. ,in, It can represent the multi-dimensional temporal feature combination of the first time point in the cervical spine splicing temporal feature vector, and can include multiple temporal feature vectors of the first time point (also called the first time step, or time step 1), such as the duration of prolonged sitting at the first time point. Motion frequency Continuous head-down time wait; This can represent the multi-dimensional temporal feature combination of the second time point in the cervical spine splicing temporal feature vector. It can represent the multi-dimensional temporal feature combination of the last time point in the cervical spine splicing temporal feature vector, and tn can represent the total number of time points contained in the cervical spine splicing temporal feature vector. This application does not limit this.

[0104] Long-term and short-term feature abstraction and iterative processing can be understood as using a temporal feature processing iterative function to process the cervical spine splicing temporal feature vector. Through iterative learning over multiple time steps, more abstract and representative features (such as fatigue accumulation trends and posture change patterns) are gradually extracted, while noise and redundant information (such as instantaneous posture fluctuations) in the original data are filtered out. During the iteration process, key temporal information (such as the cumulative effect of prolonged sitting) can be retained, and the feature representation of the current time step is dynamically updated based on the feature results of the previous time step. Finally, the target abstract memory feature vector that reflects the overall temporal pattern is output.

[0105] Optionally, the process of performing long-short-term feature abstraction and iterative processing on the cervical spine splicing temporal feature vector to obtain the target abstract memory feature vector can be seen in the following formula:

[0106]

[0107]

[0108]

[0109] in, It can be understood as short-term memory data at time step t, which can be a feature representation of time step t, generated based on previous information and current features, and can be used as one of the inputs for the next time step; It can be understood as long-term memory data at time step t, which can record key information in time series data, such as fatigue accumulation from prolonged sitting; f() can represent the process of operating through the iterative function of time series feature processing; It can represent short-term memory data at time step t-1; It can represent long-term memory data at time step t-1; it is understandable that... Both can be used as inputs to time step t to compute the state update at time step t; It can represent the multivariate temporal feature vector of time step t in the cervical spine splicing temporal feature vector; It can represent the short-term memory data at the final time step tn; This can be understood as the long-term memory data of the final time step tn; It can represent the short-term memory data of the time step preceding the final time step, i.e., time step t-1; This can represent the long-term memory data of the time step preceding the final time step, i.e., time step t-1; it is understandable that... They can all be used as inputs to the final time step tn to compute the state update of the final time step tn; It can represent the multi-terminal feature vector of the final time step tn in the cervical spine splicing temporal feature vector; It can represent the temporal feature vector of cervical spine status.

[0110] It should be noted that in the processing of cervical spine status-related data, this data is usually a continuously monitored time series. Therefore, initialization can be performed with all zeros, i.e. Both can be 0, indicating no historical state at the initial moment; in subsequent time step iterations (i.e., long-term and short-term feature abstraction iteration processing), the iteration result of the previous time step can be used as the input of the current iteration time step to maintain the continuity of temporal information such as fatigue accumulation and posture changes. For example, the state of prolonged sitting can be continuously recorded. middle.

[0111] Understandable It can be viewed as the original temporal feature sequence, such as the original numerical value of sedentary duration; This can be viewed as a feature vector after long-term and short-term feature abstraction and iteration, such as the semantic representation of "fatigue accumulation trend". Through long-term and short-term feature abstraction and iteration, and Collaborative updates enable dynamic memorization and feature extraction of time-series data. As long-term memory of the previous time step, it is the core of the long-term sequence information "remembered" in the long-short-term feature abstraction and iterative processing process; original temporal features After long-term and short-term feature abstraction and iterative processing, it is transformed into a time feature vector with semantic information. This can provide key features of the time dimension for subsequent spatiotemporal fusion.

[0112] In this embodiment, multi-dimensional temporal feature extraction and long-short-term feature abstraction and iterative processing can effectively solve the problem of time-dependent modeling of cervical spine status-related data. By extracting multi-dimensional features such as prolonged sitting time, movement frequency, and continuous head-down time and concatenating them into a temporal feature vector, the long-term memory capacity during the long-short-term feature abstraction and iterative processing can be combined to capture hidden patterns in the temporal dimensions of cervical spine fatigue accumulation and movement habits (such as linear fatigue growth caused by prolonged sitting). Compared with traditional temporal analysis methods, long-short-term feature abstraction and iterative processing can automatically filter noise and highlight key temporal features (such as fatigue peaks), enabling the generated cervical spine status temporal feature vector to more accurately reflect the dynamic changes in cervical spine health, which is beneficial for providing key temporal evidence for subsequent assessment.

[0113] In one possible implementation, the development of a target cervical spine care strategy can be achieved through a cervical spine care strategy development module within a state-detection-based cervical spine care system. Specifically, it can be achieved through the implementation steps of "abnormality type determination → strategy matching → parameter optimization → priority ranking," generating a precise target cervical spine care strategy based on personalized basic data, cervical spine abnormality early warning data, and unhealthy state assessment results. This forms a closed-loop management system for cervical spondylosis detection and care, from data-driven to personalized intervention. Specifically, a method for developing a target cervical spine care strategy based on the personalized basic data, the cervical spine abnormality early warning data, and the unhealthy state assessment results, to obtain the target cervical spine care strategy for the target detection object, includes:

[0114] D1. Determine the type of cervical spine abnormality based on the cervical spine abnormality warning data and the non-health status assessment results to obtain cervical spine abnormality type data;

[0115] D2. Based on the cervical spine abnormality type data, perform strategy matching from the target cervical spine care strategy library to obtain a matching cervical spine care strategy.

[0116] D3. Based on the cervical spine abnormality early warning data and the non-health status assessment results, the parameters of the matching cervical spine care strategy are optimized to obtain the optimized cervical spine care strategy.

[0117] D4. Based on the personalized basic data, the optimized cervical spine care strategy is prioritized to obtain the target cervical spine care strategy.

[0118] The cervical spine abnormality type data can be used to indicate the classification results of cervical spine abnormality states based on cervical spine abnormality warning data and non-health status assessment results. For example, based on cervical spine abnormality warning data, such as "orange warning: continuous sitting for 120 minutes and the curvature of the C5-C6 segment of the spine deviates by 15°", and non-health status assessment results, such as "health score of 75 points", the cervical spine abnormality type can be determined to obtain cervical spine abnormality type data such as "sit-fatigue type" or "mixed abnormality". This application does not limit this.

[0119] Optionally, the abnormality scores in the time and spatial dimensions can be calculated and then fused to obtain a fused abnormality assessment value, which can then be used to determine the type of abnormality. The time dimension abnormality score measures the cumulative degree of cervical spine abnormalities over time, focusing on indicators related to "continuous behavior" (such as duration of continuous sitting, total daily head-down time, etc.). The spatial dimension abnormality score measures the degree of abnormality in the cervical spine's physiological structure or postural space, focusing on indicators related to "physical morphology" (such as cervical curvature deviation, head tilt angle, etc.). Both dimensions of abnormality can be calculated based on the ratio between the actual monitored value and the health threshold. For example, taking the calculation of the time dimension abnormality score as an example, if the health threshold is "continuous sitting ≤ 90 minutes," and the actual monitored continuous sitting time is 135 minutes, then the time dimension abnormality score At = 135 / 90 = 1.5 (where a value > 1 indicates a time dimension abnormality, and the larger the value, the more severe the abnormality). This application does not impose any restrictions on this.

[0120] Optionally, the process of determining the type of cervical spine abnormality based on the cervical spine abnormality early warning data and the unhealthy state assessment results to obtain cervical spine abnormality type data can be found in the following formula:

[0121]

[0122] Where T can represent the fusion anomaly assessment value; It can represent the weight of the time dimension; It can represent the degree of anomaly in the time dimension; It can represent the weight of spatial dimensions; It can represent the degree of spatial dimensional anomaly.

[0123] It should be noted that the time dimension weight represents the proportion of importance of time dimension abnormalities in the comprehensive assessment, with a value ranging from 0 to 1, and can be set by clinical data or expert experience. The spatial dimension weight represents the proportion of importance of spatial dimension abnormalities in the comprehensive assessment, with a value ranging from 0 to 1, and the sum of the spatial dimension weight and the time dimension weight is 1.

[0124] By weighted fusion of anomalies in the temporal and spatial dimensions to obtain a fused anomaly assessment value, the combined impact of accumulated damage over time and spatial structural abnormalities can be quantified. This avoids the limitations of single-dimensional assessments and allows for more accurate classification of cervical spine abnormalities (such as "sedentary fatigue type," "curvature abnormal type," or "mixed type"). Specifically, if T is mainly composed of... Contributions (such as) =1.8, =0.9, and =0.5, =0.5, then T=1.35), at this time it can be judged as "sedentary fatigue type"; if T is mainly caused by Contributions (such as) =0.8, =1.7, then T=1.25), at this point it can be determined as "curvature anomaly type"; if and All are significantly >1 (e.g. =1.5, =1.4, then T=1.45), at this time it can be determined as "mixed type", and this application does not restrict it.

[0125] The target cervical spine care strategy library can be understood as a pre-stored collection of standardized care strategy templates. This library can be categorized by cervical spine abnormality type, such as strategies matching abnormal curvature, strategies matching relief from prolonged sitting, or strategies matching mixed abnormalities. Optionally, the target cervical spine care strategy library may include recommendations for different types of treatment, relief, or care plans, such as "cervical traction therapy templates," "acupoint massage templates," or "rest time planning templates," etc., and this application does not impose any limitations on this.

[0126] Matching cervical spine care strategies can be understood as matching strategy templates from a target cervical spine care strategy library based on cervical spine abnormality type data. In other words, matching cervical spine care strategies can be a rough initial cervical spine care strategy obtained through template matching. For example, for "structural abnormality type," the target cervical spine care strategy library could match: "traction therapy, once daily; acupoint massage, three times daily," but this application does not impose any limitations on this.

[0127] Optionally, the matching cervical spine care strategies can include an initial weight for each strategy. This initial weight represents the degree of matching and importance of different cervical spine care strategies with the target user's current cervical spine abnormality type. When a user is diagnosed with conditions such as "straightening of the cervical spine's physiological curvature," multiple strategies in the strategy library may be related to this type. The weight determines which strategies are prioritized. For example, for "straightening of the cervical spine's physiological curvature," the target cervical spine care strategy library might include "cervical traction training" (weight 0.8), "posture correction courses" (weight 0.7), and "heat therapy" (weight 0.5). In other words, the higher the weight, the higher the strategy's ranking in the matching results, and the higher its priority for recommendation.

[0128] For example, Table 1 below illustrates a target cervical spine care strategy library. This target cervical spine care strategy library is illustrated by including cervical spine abnormality types, recommended care strategies, basic strategy weights, and implementation suggestions, and does not constitute a limitation on this application.

[0129] Table 1

[0130]

[0131] Optimizing cervical spine care strategies can be understood as adjusting certain parameters (such as frequency and intensity) of the matching cervical spine care strategy based on the current actual situation of the target subject (i.e., cervical spine abnormality warning data and non-health status assessment results) to better fit the reality. For example, based on the "orange warning (moderate abnormality)" indicated in the cervical spine abnormality warning data, the critical cervical spine care strategy of "traction once a day" can be optimized and adjusted to "traction twice a day, 20 minutes each time".

[0132] Optionally, taking the parameters of the cervical spine care strategy, including traction time and massage intensity, as an example, the parameters of the matched cervical spine care strategy are optimized based on the cervical spine abnormality early warning data and the unhealthy state assessment results. The process of optimizing the cervical spine care strategy can be seen in the following formula:

[0133]

[0134] in, This can indicate an optimized cervical spine care strategy; It can represent the basic parameters matched in the target cervical spine care strategy library; k can represent the parameter adjustment coefficient, which can be set according to clinical data, such as 0.15; This can mean taking the maximum value between the time dimension anomaly degree and the spatial dimension anomaly degree, that is, taking the most severe anomaly dimension as the basis for adjustment; It can represent the degree of anomaly in the time dimension; It can represent the degree of spatial dimensional anomaly.

[0135] For example, if the base traction time is 20 minutes, At = 1.33, and As = 1.2, then max(At, As) = 1.33, and the optimized traction time is: =20(1+0.15*1.33)≈24 minutes. Using the above method, a quantitative correlation can be established between the severity of cervical spine abnormalities and the intensity of nursing parameters. This allows nursing strategies to be upgraded from fixed templates to dynamic adaptations, further achieving the goal of more precise interventions for more severe abnormalities. This enables dynamic adjustment of the basic parameters of nursing strategies, making nursing plans more closely aligned with the actual abnormal state, thereby improving their targeting and effectiveness.

[0136] After obtaining the optimized cervical spine care strategy, it can be further prioritized based on personalized baseline data. This means adjusting the priority order of the cervical spine care strategy according to the individual needs of the target individual, resulting in a more suitable and personalized cervical spine care strategy. For example, the feasibility of strategy implementation can be assessed based on different occupational characteristics; programmers could prioritize "desktop-executable stretching exercises." Risk tolerance can be determined based on different ages or health histories; for example, elderly individuals could prioritize addressing osteoporosis-related risks. Intervention methods can be adjusted based on different body shape parameters; for example, subjects with a BMI > 30 could avoid high-intensity traction as much as possible. This application does not impose any limitations on these aspects.

[0137] Optionally, the process of prioritizing the optimized cervical spine care strategy based on the personalized basic data to obtain the target cervical spine care strategy can be seen in the following formula:

[0138]

[0139] in, This can represent the priority score of the i-th optimization strategy. It can be understood that the higher the priority score, the higher the priority. can represent the basic weight of the strategy in the target cervical spine care strategy library; n can represent the dimension of the personalized basic data, such as if the personalized basic data includes occupation, age and exercise frequency, then its dimension can be 3; j can represent the dimension index of the personalized basic data. This can represent the weight of the j-th personalized dimension, such as occupation weight 0.5, age weight 0.3, and exercise frequency weight 0.2; It can represent the fit between the i-th strategy and the j-th personalized dimension, and the value can be 0-1.

[0140] It should be noted that, This can be a quantitative indicator measuring the degree of matching between nursing strategies and individual characteristics. The fit between the i-th strategy and the j-th personalized dimension. These values ​​can be obtained from a target cervical spine care strategy library, and their values ​​can be determined through clinical experience, behavioral data statistics, or feature mapping models. For example, a programmer's approach to "desktop stretching"... =0.9, for "traction therapy" =0.3.

[0141] The method provided by the above formula allows us to combine the fundamental value of a strategy with its individual suitability to calculate a strategy priority score. This better ensures that strategy feasibility takes precedence; for example, a programmer's "desktop stretching" strategy... (Job suitability) is extremely high, even with its base weight. It is ranked lower than "traction therapy", thus receiving a higher priority score. The recommendation rate is still higher, avoiding strategies that are theoretically effective but difficult to implement in practice; furthermore, it can better highlight the impact of individual differences on strategies, for example, the same "swimming rehabilitation" strategy may have different effects on "sports enthusiasts" ( =0.8) has a much higher priority than "people with physical disabilities" ( =0.1), through (Personalized dimensional weighting) amplifies the impact of key features (such as "athletic ability"), achieving truly personalized ranking; furthermore, it can better balance the effectiveness and adaptability of the strategy, for example, in the formula (Base weighting) ensures that more effective medical strategies (such as traction therapy) are not completely ignored, while This allows for adjustments to the prioritization of care in individual scenarios, avoiding mismatches caused by a purely results-oriented approach; it also provides clear guidance for the implementation of subsequent care plans, such as scoring. The order of importance directly corresponds to the recommended strategy (e.g., If strategy 1 is selected first, then users or medical staff can obtain the most suitable intervention plan more quickly without manual screening.

[0142] In this embodiment, a data-driven process is realized, from precise localization of cervical spine abnormalities to dynamic adaptation of cervical spine care strategies. This ensures the targeted nature of cervical spine care strategies through cervical spine abnormality early warning data and unhealthy state assessment results (e.g., traction can be prioritized for abnormal curvature). It also enhances the feasibility of cervical spine care strategies by leveraging personalized basic data (e.g., desktop stretching can be adapted for programmers). This forms a closed loop of "abnormality identification - scientific matching - precise adjustment - orderly execution", which can significantly improve the efficiency and intervention effect of cervical spine care.

[0143] For example, in this application embodiment, a cervical spine care device integrating dynamic support and massage is provided, aiming to create a device that integrates dynamic support and precise massage to effectively relieve cervical spine pain. Optionally, the cervical spine care device integrating dynamic support and massage may include, but is not limited to, an adaptive dynamic support system, a precise acupoint massage module, and a health data monitoring and feedback module.

[0144] The adaptive dynamic support system is equipped with multiple pressure sensors distributed at key points in contact with the neck (similar to the functions of the data acquisition module, spatiotemporal feature fusion module, and cervical spine care strategy formulation module in the aforementioned embodiments). When the user's head posture changes, the sensors can quickly capture the pressure changes and transmit the information to the microprocessor in the system. After analysis, the microprocessor can adjust the angle and height of the support structure in real time to always provide the most suitable support for the cervical spine. In addition, the support material combines memory foam and elastic silicone, which can conform to the curve of the neck and has good support and comfort.

[0145] The precision acupoint massage module utilizes artificial intelligence (AI) technology, based on extensive neck acupoint data and massage techniques, and incorporates multiple massage programs (similar to the cervical spine care strategy formulation module in the aforementioned embodiment). The massage heads at the front of the device are wrapped in flexible material and driven by a motor to simulate the pushing, kneading, and pressing techniques of a professional massage therapist, precisely stimulating multiple acupoints related to cervical spine health, such as Fengchi and Tianzhu acupoints. Simultaneously, the massage intensity and frequency can be adjusted according to the user's needs and feedback.

[0146] The health data monitoring and feedback module can be a built-in biosensor that can monitor the user's neck muscle tension, cervical spine range of motion, and other health data in real time, and transmit the data to a companion mobile application (similar to the functions of the cervical spine health assessment module, cervical spine abnormality early warning module, and cervical spine care strategy formulation module in the aforementioned embodiments). This application can not only analyze the data and generate personalized health reports, but also intelligently recommend suitable massage and support programs based on the user's health condition.

[0147] This dynamic support and massage integrated cervical spine care device is suitable for various groups requiring cervical spine care, including office workers and the elderly. Office workers can wear it while working to reduce cervical spine pressure caused by prolonged head-down posture; drivers who frequently drive can enjoy comfortable cervical spine care while driving; and the elderly can use it at home to relieve pain caused by degenerative cervical spine diseases, etc. This application does not impose any limitations on these uses.

[0148] For examples consistent with the above embodiments, please refer to... Figure 3 , Figure 3 This is a schematic diagram of the structure of a terminal provided in an embodiment of this application, such as... Figure 3 As shown, it includes a processor, an input device, an output device, and a memory, which are interconnected. The memory is used to store a computer program, which includes program instructions. The processor is configured to call the program instructions. The program includes instructions for performing the following steps.

[0149] The system acquires personalized basic data of the target detection object, as well as cervical spine state-related data and cervical spine posture-related images of the target detection object collected by the sensor network.

[0150] Based on the personalized basic data, the cervical spine state-related data, and the cervical spine posture-related images, spatiotemporal feature fusion processing is performed to obtain the spatiotemporal matrix of the cervical spine state of the target detection object;

[0151] Based on the personalized basic data and the cervical spine status spatiotemporal matrix, a cervical spine health assessment is performed to obtain the cervical spine health status assessment result of the target detection object.

[0152] When the non-health status assessment result in the cervical spine health status assessment results meets the preset conditions, cervical spine abnormality early warning data of the target detection object is generated.

[0153] Based on the personalized basic data, the cervical spine abnormality early warning data, and the unhealthy state assessment results, a target cervical spine care strategy is formulated to obtain the target cervical spine care strategy for the target detection object.

[0154] The above mainly describes the solutions of the embodiments of this application from the perspective of the method execution process. It is understood that, in order to achieve the above functions, the terminal includes the corresponding hardware structure and / or software modules for executing each function. Those skilled in the art should readily recognize that, in conjunction with the units and algorithm steps of the various examples described in the embodiments provided herein, this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed in hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0155] This application embodiment can divide the terminal into functional units according to the above method example. For example, each function can be divided into a separate functional unit, or two or more functions can be integrated into one processing unit. The integrated unit can be implemented in hardware or as a software functional unit. It should be noted that the unit division in this application embodiment is illustrative and only represents one logical functional division. In actual implementation, there may be other division methods.

[0156] For those consistent with the above, please refer to Figure 4 , Figure 4 This application provides a schematic diagram of a cervical spine care device based on state detection. Figure 4 As shown, the device includes:

[0157] The acquisition unit 101 is used to acquire personalized basic data of the target detection object, as well as cervical spine state-related data and cervical spine posture-related images of the target detection object collected by the sensor network.

[0158] The first processing unit 102 is used to perform spatiotemporal feature fusion processing based on the personalized basic data, the cervical spine state-related data and the cervical spine posture-related image to obtain the spatiotemporal matrix of the cervical spine state of the target detection object.

[0159] The second processing unit 103 is used to perform cervical spine health assessment processing based on the personalized basic data and the cervical spine status spatiotemporal matrix to obtain the cervical spine health status assessment result of the target detection object.

[0160] The third processing unit 104 is used to generate cervical spine abnormality warning data of the target detection object when the non-healthy state assessment result in the cervical spine health status assessment result meets the preset conditions.

[0161] The fourth processing unit 105 is used to formulate a target cervical spine care strategy based on the personalized basic data, the cervical spine abnormality early warning data and the unhealthy state assessment results, thereby obtaining the target cervical spine care strategy for the target detection object.

[0162] In one possible implementation, the first processing unit 102 is configured to perform spatiotemporal feature fusion processing based on the personalized basic data, the cervical spine state-related data, and the cervical spine posture-related image to obtain the spatiotemporal matrix of the cervical spine state of the target detection object, specifically for:

[0163] The personalized basic data is processed by personalized weight calculation to obtain a personalized weight vector;

[0164] The temporal features of the cervical spine state-related data are extracted to obtain the temporal feature vector of the cervical spine state.

[0165] State space features are extracted from the cervical spine state-related data to obtain the cervical spine state space feature vector;

[0166] The posture angle features of the cervical spine posture-related images are extracted to obtain the cervical spine posture angle feature vector;

[0167] The cervical spine state space feature vector and the cervical spine posture angle feature vector are fused to obtain the cervical spine fused space feature vector.

[0168] The personalized weight vector, the cervical spine state temporal feature vector, and the cervical spine fusion spatial feature vector are spatiotemporally fused to obtain the spatiotemporal matrix of the cervical spine state of the target detection object.

[0169] In one possible implementation, the first processing unit 102 is configured to perform personalized weight calculation processing on the personalized basic data to obtain a personalized weight vector, specifically for:

[0170] The personalized basic data is quantized to obtain a personalized basic feature vector;

[0171] The personalized basic feature vector is subjected to a linear transformation to obtain a personalized linear feature vector;

[0172] The personalized linear feature vector is processed by feature importance mapping matrix to obtain personalized mapped feature vector;

[0173] The personalized mapping feature vector is normalized to obtain the personalized weight vector.

[0174] In one possible implementation, the first processing unit 102 is configured to extract temporal features from the cervical spine state-related data to obtain a cervical spine state temporal feature vector, specifically for:

[0175] Multi-dimensional temporal feature extraction processing is performed on the cervical spine state-related data to obtain a reference temporal feature vector set;

[0176] Each reference time-series feature vector in the reference time-series feature vector set is concatenated to obtain the cervical spine concatenated time-series feature vector.

[0177] The cervical spine splicing temporal feature vector is subjected to long-term and short-term feature abstraction and iteration processing to obtain the target abstract memory feature vector;

[0178] The target abstract memory feature vector is used as the cervical spine state temporal feature vector.

[0179] In one possible implementation, the fourth processing unit 105 is configured to formulate a target cervical spine care strategy based on the personalized basic data, the cervical spine abnormality early warning data, and the unhealthy state assessment results, thereby obtaining the target cervical spine care strategy for the target detection object, specifically for:

[0180] Based on the cervical spine abnormality early warning data and the non-health status assessment results, the cervical spine abnormality type is determined to obtain cervical spine abnormality type data.

[0181] Based on the cervical spine abnormality type data, a matching cervical spine care strategy is obtained by matching the strategy from the target cervical spine care strategy library.

[0182] Based on the cervical spine abnormality early warning data and the non-health status assessment results, the parameters of the matching cervical spine care strategy are optimized to obtain an optimized cervical spine care strategy.

[0183] Based on the personalized basic data, the optimization cervical spine care strategy is prioritized to obtain the target cervical spine care strategy.

[0184] This application also provides a computer storage medium storing a computer program for electronic data interchange, which causes a computer to perform some or all of the steps of any of the state-detection-based cervical spine care methods described in the above method embodiments.

[0185] This application also provides a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program that causes a computer to perform some or all of the steps of any of the state-detection-based cervical spine care methods described in the above method embodiments.

[0186] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0187] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0188] In the several embodiments provided in this application, it should be understood that the disclosed apparatus can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical or other forms.

[0189] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0190] Furthermore, the functional units in the various embodiments of the application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software program module.

[0191] If the integrated unit is implemented as a software program module and sold or used as an independent product, it can be stored in a computer-readable storage device (CMD). Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned memory includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0192] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage device, which may include: a flash drive, a read-only memory, a random access memory, a magnetic disk, or an optical disk, etc.

[0193] The embodiments of this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A cervical spine care method based on state detection, characterized in that, The method includes: The system acquires personalized basic data of the target detection object, as well as cervical spine state-related data and cervical spine posture-related images of the target detection object collected by the sensor network. Based on the personalized basic data, the cervical spine state-related data, and the cervical spine posture-related images, spatiotemporal feature fusion processing is performed to obtain the spatiotemporal matrix of the cervical spine state of the target detection object; Based on the personalized basic data and the cervical spine status spatiotemporal matrix, a cervical spine health assessment is performed to obtain the cervical spine health status assessment result of the target detection object. When the non-health status assessment result in the cervical spine health status assessment results meets the preset conditions, cervical spine abnormality early warning data of the target detection object is generated. Based on the personalized basic data, the cervical spine abnormality early warning data, and the unhealthy state assessment results, a target cervical spine care strategy is formulated to obtain the target cervical spine care strategy for the target detection object. The step of performing spatiotemporal feature fusion processing based on the personalized basic data, the cervical spine state-related data, and the cervical spine posture-related images to obtain the spatiotemporal matrix of the cervical spine state of the target detection object includes: The personalized basic data is quantized to obtain a personalized basic feature vector; The personalized basic feature vector is subjected to a linear transformation to obtain a personalized linear feature vector; The personalized linear feature vector is processed by feature importance mapping matrix to obtain personalized mapped feature vector; The personalized mapping feature vector is normalized to obtain the personalized weight vector; The temporal features of the cervical spine state-related data are extracted to obtain the temporal feature vector of the cervical spine state. State-space feature extraction is performed on the cervical spine state-related data to obtain a cervical spine state-space feature vector; the state-space feature extraction includes the calculation of spinal curvature; the calculation of spinal curvature is performed by acquiring the spinal curvature data of segments C3-C7 of the target object through a posture sensor. ,in, The actual spinal curvature value of the third cervical segment. The actual spinal curvature value of the fourth cervical segment. The actual spinal curvature value of the 5th cervical segment. The actual spinal curvature value of the 6th cervical segment. The actual spinal curvature value of the 7th cervical segment; and the difference from the normal spinal curvature is calculated sequentially. ,in, This represents the absolute difference between the actual curvature and the normal curvature of the i-th cervical vertebra. This represents the actual curvature of the i-th cervical vertebra. Indicates normal curvature; The posture angle features of the cervical spine posture-related images are extracted to obtain the cervical spine posture angle feature vector; The cervical spine state space feature vector and the cervical spine posture angle feature vector are fused to obtain the cervical spine fused space feature vector. The personalized weight vector, the cervical spine state temporal feature vector, and the cervical spine fusion spatial feature vector are subjected to spatiotemporal fusion processing to obtain the spatiotemporal matrix of the cervical spine state of the target detection object.

2. The cervical spine care method based on state detection according to claim 1, characterized in that, The step of extracting temporal features from the cervical spine state-related data to obtain a cervical spine state temporal feature vector includes: Multi-dimensional temporal feature extraction processing is performed on the cervical spine state-related data to obtain a reference temporal feature vector set; Each reference time-series feature vector in the reference time-series feature vector set is concatenated to obtain the cervical spine concatenated time-series feature vector. The cervical spine splicing temporal feature vector is subjected to long-term and short-term feature abstraction and iteration processing to obtain the target abstract memory feature vector; The target abstract memory feature vector is used as the cervical spine state temporal feature vector.

3. The cervical spine care method based on state detection according to claim 1 or 2, characterized in that, The step of formulating a target cervical spine care strategy based on the personalized basic data, the cervical spine abnormality early warning data, and the unhealthy state assessment results, to obtain the target cervical spine care strategy for the target detection subject, includes: Based on the cervical spine abnormality early warning data and the non-health status assessment results, the cervical spine abnormality type is determined to obtain cervical spine abnormality type data. Based on the cervical spine abnormality type data, a matching cervical spine care strategy is obtained by matching the strategy from the target cervical spine care strategy library. Based on the cervical spine abnormality early warning data and the non-health status assessment results, the parameters of the matching cervical spine care strategy are optimized to obtain an optimized cervical spine care strategy. Based on the personalized basic data, the optimization cervical spine care strategy is prioritized to obtain the target cervical spine care strategy.

4. A cervical spine care device based on state detection, characterized in that, The device includes: The acquisition unit is used to acquire personalized basic data of the target detection object, as well as cervical spine state-related data and cervical spine posture-related images of the target detection object collected by the sensor network. The first processing unit is used to perform spatiotemporal feature fusion processing based on the personalized basic data, the cervical spine state-related data and the cervical spine posture-related image to obtain the spatiotemporal matrix of the cervical spine state of the target detection object. The second processing unit is used to perform cervical spine health assessment processing based on the personalized basic data and the cervical spine status spatiotemporal matrix to obtain the cervical spine health status assessment result of the target detection object. The third processing unit is used to generate cervical spine abnormality warning data of the target detection object when the non-healthy status assessment result in the cervical spine health status assessment result meets the preset conditions. The fourth processing unit is used to formulate a target cervical spine care strategy based on the personalized basic data, the cervical spine abnormality early warning data and the unhealthy state assessment results, so as to obtain the target cervical spine care strategy for the target detection object. The first processing unit is configured to perform spatiotemporal feature fusion processing based on the personalized basic data, the cervical spine state-related data, and the cervical spine posture-related images to obtain the spatiotemporal matrix of the cervical spine state of the target detection object, specifically for: The personalized basic data is quantized to obtain a personalized basic feature vector; The personalized basic feature vector is subjected to a linear transformation to obtain a personalized linear feature vector; The personalized linear feature vector is processed by feature importance mapping matrix to obtain personalized mapped feature vector; The personalized mapping feature vector is normalized to obtain the personalized weight vector; The temporal features of the cervical spine state-related data are extracted to obtain the temporal feature vector of the cervical spine state. State-space feature extraction is performed on the cervical spine state-related data to obtain a cervical spine state-space feature vector; the state-space feature extraction includes the calculation of spinal curvature; the calculation of spinal curvature is performed by acquiring the spinal curvature data of segments C3-C7 of the target object through a posture sensor. ,in, The actual spinal curvature value of the third cervical segment. The actual spinal curvature value of the fourth cervical segment. The actual spinal curvature value of the 5th cervical segment. The actual spinal curvature value of the 6th cervical segment. The actual spinal curvature value of the 7th cervical segment; and the difference from the normal spinal curvature is calculated sequentially. ,in, This represents the absolute difference between the actual curvature and the normal curvature of the i-th cervical vertebra. This represents the actual curvature of the i-th cervical vertebra. Indicates normal curvature; The posture angle features of the cervical spine posture-related images are extracted to obtain the cervical spine posture angle feature vector; The cervical spine state space feature vector and the cervical spine posture angle feature vector are fused to obtain the cervical spine fused space feature vector. The personalized weight vector, the cervical spine state temporal feature vector, and the cervical spine fusion spatial feature vector are subjected to spatiotemporal fusion processing to obtain the spatiotemporal matrix of the cervical spine state of the target detection object.

5. A terminal, characterized in that, The device includes a processor, an input device, an output device, and a memory, which are interconnected. The memory stores a computer program, which includes program instructions. The processor is configured to invoke the program instructions to execute the cervical spine care method based on state detection as described in any one of claims 1-3.

6. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, the computer program including program instructions that, when executed by a processor, cause the processor to perform the cervical spine care method based on state detection as described in any one of claims 1-3.

Images