A method of rehabilitation exercise assessment

Abstract

The application provides a rehabilitation exercise evaluation method, and belongs to the technical field of intelligent medical software, and has the technical points that the method comprises the following steps: in a first stage, whether the action during rehabilitation exercise is standard is evaluated; and in a second stage, a muscle state is evaluated. The rehabilitation exercise evaluation method can more accurately evaluate the rehabilitation progress and training effect of a patient.

Description

Technical Field

[0001] This invention belongs to the technical field of intelligent medical software, and specifically relates to a method for assessing rehabilitation exercises. Background Technology

[0002] The goal of rehabilitation exercise therapy is to help patients restore their physical function to the greatest extent possible and improve their quality of life. Therefore, it is very important to assess the effectiveness of the rehabilitation exercise process.

[0003] The combination of muscle electrical impedance tomography (M-EIT) with visual human motion capture methods provides a novel solution for human motion capture and analysis.

[0004] M-EIT technology can monitor changes in tissue physiological characteristics in real time, while visual human body capture methods focus on acquiring human movement and posture information through visual sensors. These two technologies combine the internal physiological information provided by EIT with the external morphological information provided by visual capture to obtain a more comprehensive analysis of human movement and physiological state.

[0005] In rehabilitation training, M-EIT can monitor muscle activation and deep tissue changes, while visual capture can provide the trajectory and postural changes of limb movements. Combining these two types of data allows for a more accurate assessment of the patient's rehabilitation progress and training effectiveness. This combination provides a new, multi-dimensional perspective for capturing and analyzing human movement, contributing to more precise human motion monitoring and evaluation. Summary of the Invention

[0006] The purpose of this invention is to solve the problems existing in the prior art and provide a method for evaluating rehabilitation exercise standards.

[0007] Another objective of this invention is to provide a method for assessing muscle condition during rehabilitation exercises.

[0008] Another objective of this invention is to provide a method for assessing rehabilitation exercises.

[0009] Another object of the present invention is to provide a storage medium.

[0010] The technical solution of this application is: A rehabilitation exercise movement standard assessment method evaluates the movement by comparing the directions of any two adjacent bones representing a pre-set standard movement and the joint angle between those two bones with the directions of any two adjacent bones representing a measured movement and the joint angle between those two bones.

[0011] Furthermore, the directions of any two adjacent bones and the joint between the two bones are: the direction of the left upper arm and the left forearm, and the angle of the left elbow joint. And / or, the direction of the right upper arm and right forearm, and the angle of the right elbow joint; And / or, the direction of the left thigh and left calf, and the angle of the left knee joint; And / or, the direction of the right thigh, the right calf, and the angle of the right knee joint.

[0012] Furthermore, the method for evaluating any two adjacent bones—the first bone, the second bone, and the joint angle between them—includes the following steps: S100, obtain M and P in the same coordinate system; , , Where M is the coordinate matrix of the three attitude sensors representing the standard action; (x) 标准-1 y 标准-1 , z 标准-1 ), (x 标准-2 y 标准-2 , z 标准-2 ), (x 标准-3 y 标准-3 , z 标准-3 The following represent the three-dimensional coordinates of the posture sensors at the first bone, joint, and second bone, respectively, representing the standard movement. (For example: reading the three-dimensional coordinates of the left shoulder joint posture sensor 101 (i.e., the first bone), the left elbow joint posture sensor 102 (i.e., the joint), and the left wrist joint posture sensor 103 (i.e., the second bone). Essentially, this requires reading the coordinates of the posture sensors at three adjacent joints.) P is the coordinate matrix of the human body's center of mass characteristic points during standard movements (x 质心-标准 y 质心-标准， z 质心-标准 () represents the three-dimensional coordinates of the human body's center of mass characteristic point during standard movements; S200, obtain R and P in the same coordinate system S ; , , Where R is the coordinate matrix of the three attitude sensors representing the measured action; (x 实测-1 y 实测-1 , z 实测-1 ), (x 实测-2 y 实测-2 , z 实测-2 ), (x 实测-3 y 实测-3 , z 实测-3) represent the three-dimensional coordinates of the posture sensors at the first bone, joint, and second bone, respectively, representing the actual measured movement; Among them, P S The coordinate matrix representing the human body's center of mass during the measured action; (x 质心-实测 y 质心-实测， z 质心-实测 () represents the three-dimensional coordinates of the human body's center of mass feature point during the measured action; S300, solve for the relative coordinate matrix M' of the three attitude sensors when representing the standard action, and the relative coordinate matrix R' of the three attitude sensors when representing the measured action; , , S400, solve for the real-time registration matrix R'', which is solved using steps S401 to S404; S401, Solve for the cross-covariance matrix H using M' and R': H=M' T ·R', where M' T It is the transpose of M'; S402, Perform SVD decomposition on H: H = V·S·W T V is a 3×3 orthogonal matrix, S is a 3×3 diagonal matrix, and W is a 3×3 orthogonal matrix; S403, Solve for the optimal rotation matrix U: S4031, U=W·V T V T It is the transpose of V; S4032, if det(U)>0, U is not modified; if det(U)<0, the third column of W is inverted and then ANDed with V. T The product of these is the result of U; S404, R''=R'·U; R'' is expressed as: , Where, x 实测-1 ''、x 实测-2 ''、x 实测-3 ''、y 实测-1 ''、y 实测-2 ''、y 实测-3 ''、z 实测-1 ''、z 实测-2 ''、z 实测-3 Elements that are R; S500, solve for the direction vectors V of the first and second bones when representing the standard movement. std1 V std2 and joint angle α std ; V std1 = (x 标准-1 -x 标准-2 y 标准-1 -y 标准-2 , z 标准-1 -z 标准-2 ); V std2 = (x 标准-3 -x 标准-2 y 标准-3 -y 标准-2 , z 标准-3 -z 标准-2 ); α std =arccos((V std1 ·V std2 ) / (|V std1 |×|V std2 |)) S600, solve for the direction vectors V of the first and second skeletons when representing the measured motion. real1 V real2 and joint angle α real ; V real1 = (x 实测-1 ''-x 实测-2 '', y 实测-1 ''-y 实测-2 '', z 实测-1 ''-z 实测-2 ''); V real2 = (x 实测-3 ''-x 实测-2 '', y 实测-3 ''-y 实测-2 '', z 实测-3 ''-z 实测-2 ''); α real =arccos((V real1 ·V real2 ) / (|V real1 |×|V real2 |)) S700, Logical Comparison: If |α std -α real |≤δ 阈值 β1≤β 1阈值 β2≤β 2阈值 The action meets the requirements; Otherwise, the actions do not match; Wherein, the angle β1 between the first skeleton of the measured movement and the first skeleton of the standard movement is β1 = acrcos((Vreal1 ·V std1 ) / |V real1 |×|V std1 |); Among them, the angle β between the second skeleton of the measured movement and the second skeleton of the standard movement. 2= =acrcos((V real2 ·V std2 ) / |V real2 |×|V std2 |); Where, δ 阈值 This represents the threshold for joint angle motion matching between the first and second bones, in degrees. β 1阈值 This represents the threshold angle for the change in the direction of movement of the first bone, in degrees. β 2阈值 This represents the threshold angle for the change in the direction of movement of the second bone, expressed in degrees.

[0013] A method for assessing muscle condition during rehabilitation exercises includes the following steps: S100, obtain the time-difference relative conductivity matrix of the muscle to be tested; S200, solve for the average value Δσ of the time-difference relative conductivity matrix. mean According to △σ mean To determine the muscle mass status: When Δσ mean =0 indicates that the muscle is in an inactive state; when Δσ = 0, it indicates that the muscle is inactive; mean ≠0 indicates that the muscle is in an activated state: (1) △σ mean ≤a 阈值 This indicates a relatively small total muscle mass; (2) a 阈值 <△σ mean ≤b 阈值 This indicates that the muscle mass is within the normal range; (3)b 阈值 <△σ mean This indicates a high muscle mass; among them, a 阈值 b 阈值 It is a threshold.

[0014] A method for assessing muscle condition during rehabilitation exercises includes the following steps: Step A: Acquire an image of the muscle and extract its contour line; the region within the contour line is the EIT simulation model region; read the boundary voltage difference matrix ΔV of the muscle and solve for the unit conductivity change matrix Δσ. LBP Determine whether there are dynamic areas of interest; △σ LBP =J T· △V; where J is the conductivity sensitivity matrix and △V is the time-difference boundary voltage difference matrix of the muscle to be detected; The method for determining whether there is a dynamic region of interest is: if the matrix Δσ LBP The absolute value of any element in the array is greater than Δσ. t When this occurs, the muscle has a dynamic region of interest; △σ t For the threshold; If a dynamic region of interest exists in the muscle, proceed to step B; otherwise, do not proceed to step B. Step B: Perform targeted high-precision EIT reconstruction and analysis on the muscle: output the reconstructed conductivity distribution image; solve the time-difference relative conductivity matrix of the muscle through the boundary voltage difference matrix ΔV, and then output the reconstructed conductivity distribution image; Step C: Based on the time difference relative conductivity matrix obtained in Step B, calculate the average value Δσ of the time difference relative conductivity matrix. mean According to △σ mean To determine the muscle mass status: When Δσ mean =0 indicates that the muscle is in an inactive state; when Δσ = 0, it indicates that the muscle is inactive; mean ≠0 indicates that the muscle is in an activated state: (1) △σ mean ≤a 阈值 This indicates a relatively small total muscle mass; (2) a 阈值 <△σ mean ≤b 阈值 This indicates that the muscle mass is within the normal range; (3)b 阈值 <△σ mean This indicates a high muscle mass; among them, a 阈值 b 阈值 It is a threshold.

[0015] A method for assessing muscle condition during rehabilitation exercises includes the following steps: S100, Acquire impedance spectrum data; The impedance analyzer, controlled by a computer, injects a current with a frequency range of f1 and f2 into the human body from a pair of electrodes, while simultaneously measuring the impedance data of the other adjacent electrode pairs. The impedance analyzer acquires impedance spectrum data once. S200, using impedance spectral data to calculate the spatial average phase angle Δφ of the muscle. mean Through △φ mea To assess the muscle's sarcoplasmic state; When △φ mean =0 indicates that the muscle is in an inactive state; when Δφ mean ≠0, indicating that the muscle is in an activated state: (1) △φ mean ≤u 阈值 This indicates that the health and integrity of muscle cells have been affected; (2) u 阈值 <△φmean ≤v 阈值 This indicates that the muscle cells are healthy; (3) v 阈值 <△φ mean This indicates that muscle cells have been enhanced; among them, u 阈值 v 阈值 It is a threshold.

[0016] A method for assessing muscle condition during rehabilitation exercises includes the following steps: Step A: Acquire an image of the muscle and extract its contour line; the region within the contour line is the EIT simulation model region; read the boundary voltage difference matrix ΔV of the muscle and solve for the unit conductivity change matrix Δσ. LBP Determine whether there are dynamic areas of interest; △σ LBP =J T · △V; where J is the conductivity sensitivity matrix and △V is the time-difference boundary voltage difference matrix of the muscle to be tested; The method for determining whether there is a dynamic region of interest is: if the matrix Δσ LBP The absolute value of any element in the array is greater than Δσ. t When this occurs, the muscle has a dynamic region of interest; △σ t For the threshold; If a dynamic region of interest exists in the muscle, proceed to step B; otherwise, do not proceed to step B. Step B: Perform targeted high-precision EIT reconstruction and analysis on the muscle: output the reconstructed conductivity distribution image; solve the time-difference relative conductivity matrix of the muscle through the boundary voltage difference matrix ΔV, and then output the reconstructed conductivity distribution image; Simultaneously, impedance spectrum data is acquired; a computer-controlled impedance analyzer injects current with frequencies f1 and f2 into the human body from a pair of electrodes, while simultaneously measuring the impedance data of the remaining adjacent electrode pairs, and the impedance analyzer acquires impedance spectrum data once; the spatial average phase angle Δφ of the muscle is then calculated from the impedance spectrum data. mean Through △φ mea To assess the muscle's sarcoplasmic state; Step C: Based on the time difference relative conductivity matrix obtained in Step B, calculate the average value Δσ of the time difference relative conductivity matrix. mean According to △σ mean To determine the muscle mass status: When Δσ mean =0 indicates that the muscle is in an inactive state; When Δσ mean ≠0 indicates that the muscle is in an activated state: (1) △σ mean ≤a 阈值 This indicates a relatively small total muscle mass; (2) a 阈值 <△σ mean≤b 阈值 This indicates that the muscle mass is within the normal range; (3)b 阈值 <△σ mean This indicates a high muscle mass; among them, a 阈值 b 阈值 It is a threshold; Through △φ mea To assess the muscle's sarcoplasmic state; When △φ mean =0 indicates that the muscle is in an inactive state; When △φ mean ≠0, indicating that the muscle is in an activated state: (1) △φ mean ≤u 阈值 This indicates that the health and integrity of muscle cells have been affected; (2) u 阈值 <△φ mean ≤v 阈值 This indicates that the muscle cells are healthy; (3) v 阈值 <△φ mean This indicates that muscle cells have been enhanced; among them, u 阈值 v 阈值 It is a threshold.

[0017] Furthermore, △σ mean ≤a 阈值 , △φ mean <u 阈值 This indicates that both muscle quality and cell count are poor, requiring stronger intervention.

[0018] A rehabilitation exercise assessment method includes the following steps: In the first stage, the aforementioned rehabilitation exercise movement standard assessment method is used to evaluate whether the movements during rehabilitation exercises are standard: if the movements meet the requirements, the second stage is carried out; otherwise, the second stage is not carried out. In the second stage, the muscle status is assessed using the aforementioned rehabilitation exercise muscle status assessment method.

[0019] A storage medium storing a program capable of performing the aforementioned rehabilitation exercise standard assessment method.

[0020] A storage medium storing a program capable of performing the aforementioned rehabilitation exercise muscle status assessment method.

[0021] A storage medium storing a program capable of performing the aforementioned rehabilitation exercise assessment method.

[0022] The advantages of the technical solution of this invention are mainly reflected in: First, the fundamental concept of this application is a rehabilitation exercise assessment method that requires evaluation from two aspects: whether the rehabilitation exercise movements are standard and the state of the muscles during the rehabilitation exercise. The rehabilitation exercise is assessed jointly from these two aspects.

[0023] Second, the second basic concept of this application is to evaluate the accuracy of the rehabilitation movement by comparing the direction of any two adjacent bones representing a pre-set standard movement and the joint angle between the two bones with the direction of any two adjacent bones representing a measured movement and the joint angle between the two bones.

[0024] Third, the third fundamental concept of this application is that it uses time difference conductivity to determine the muscle state during rehabilitation exercises.

[0025] Fourth, to conserve computing power, two imaging modes are used when assessing the muscle state during rehabilitation exercises. The initial mode calculates Δσ. LBP To save computing power. When Δσ LBP When an ROI is identified, the second mode is required: precise calculation of the time-difference relative conductivity matrix (precise imaging). Based on the time-difference relative conductivity matrix obtained in the second precise mode, the average value of the time-difference relative conductivity matrix is ​​calculated, and the muscle state is determined by the average value.

[0026] Fourth, it should be noted that the EIT imaging method of this application is also applicable to imaging methods of organs such as the brain (two-stage imaging method). Attached Figure Description

[0027] The present application will be further described in detail below with reference to the embodiments in the accompanying drawings, but this does not constitute any limitation on the present application.

[0028] Figure 1 This is a schematic diagram of the muscle electrical impedance detection component of this application.

[0029] Figure 2 This is a schematic diagram of the muscle electrical impedance detection component of this application from another perspective.

[0030] Figure 3 This is a schematic diagram of the attitude sensor assembly of this application. Detailed Implementation

[0031] The objectives, advantages, and features of this invention will be explained through the following non-limiting description of preferred embodiments. These embodiments are merely typical examples of applying the technical solutions of this invention, and all technical solutions formed by equivalent substitutions or equivalent transformations fall within the scope of protection claimed by this invention.

[0032] <Example 1: A visual human body capture device for muscle electrical impedance tomography> A visual human body capture device for muscle electrical impedance tomography, comprising: a, Muscle electrical impedance detection component 1, which is used to acquire electrical signal data of the whole body muscles of the human body under test; The muscle impedance detection component 1 includes: a right upper arm sensor component 2, a left upper arm sensor component 3, a right lower arm sensor component 4, a left lower arm sensor component 5, a right thigh sensor component 6, a left thigh sensor component 7, a right calf sensor component 8, a left calf sensor component 9, a chest cavity sensor component 10, an abdominal muscle sensor component 11, and a back muscle sensor component 12; the aforementioned sensor components 2 to 10 are wrapped around the corresponding muscles on the subject's body, and the abdominal muscle sensor component 11 and the back muscle sensor component 12 are respectively attached to the abdominal and back muscles of the subject's body; b, Visual camera 13, which is used to capture human motion posture data; c. A posture sensor assembly 100, which works in conjunction with a vision camera 13 to acquire human motion information; The posture sensor assembly 100 includes: a left shoulder joint posture sensor 101, a left elbow joint posture sensor 102, a left wrist joint posture sensor 103, a left hip joint posture sensor 104, a left knee joint posture sensor 105, a left ankle joint posture sensor 106; a right shoulder joint posture sensor 107, a right elbow joint posture sensor 108, a right wrist joint posture sensor 109, a right hip joint posture sensor 110, a right knee joint posture sensor 111, and a right ankle joint posture sensor 112.

[0033] <Example 2: A Method for Assessing Rehabilitation Exercise> A rehabilitation exercise assessment method includes the following steps: The first stage assesses whether the movements during rehabilitation exercises are standard: if the movements meet the requirements, the second stage is carried out; otherwise, the second stage is not carried out. That is, standard movements are a prerequisite for judging muscle activation. Under non-standard movements, the assessment of muscle status is not meaningful. The second stage involves assessing the muscle condition.

[0034] <Example 3: A Method for Evaluating Standardized Rehabilitation Exercise Movements> A standard assessment method for rehabilitation exercise movements includes the following steps: S100, obtain the pre-set rehabilitation standard movement dataset Q and the coordinate matrix of human centroid feature points (e.g., the coordinates of the middle part of the navel) within the time period 0 to T. T ; Q=(M0, M1, M2, ..., Mt， ...M T ); M t Let be the coordinate matrix of N attitude sensors at any time t, that is, the three-dimensional coordinate matrix of the attitude sensors corresponding to the standard action at time t (the three-dimensional coordinate matrix of the attitude sensors is actually the three-dimensional coordinate matrix of the feature points of each attitude sensor, which can characterize the position of the attitude sensors). M t P t The coordinates are obtained in the same coordinate system; S200, Obtain the measured motion dataset U and the measured human centroid feature point coordinate matrix P within the time period 0 to T. s1 ~P sT ; U = (R0, R1, R2, ..., R) t， ...R T ); R t The three-dimensional coordinate matrix of each posture sensor of the human body under test at any time t is obtained by acquiring the human body under test through a vision camera. This matrix is ​​called the coordinate matrix of the posture sensor of the N actual actions at any time t. R t P St The coordinates are obtained from the same coordinate system; S300, solve for M0 ’ ~M T ’ R0 ’ ~R T ’ ; Geometric feature-based posture structure matching and determination: In order to achieve accurate judgment of the user's posture and eliminate matching errors caused by different absolute positions or orientations in space, this system adopts a geometric feature-based structure matching method.

[0035] M t =(C 标准t-1 C 标准t-2 C 标准t-I , ...C 标准t-N ) T C 标准t-i The coordinate matrix is ​​a preset value for the i-th attitude sensor at any time t; N is the total number of attitude sensors. M t ’ =(C 标准t-1 -P t C 标准t-2 -P t C 标准t-i -P t， …C标准t-N -P t ) T M t To M t ’ Essentially, it is the relative coordinate matrix of N attitude sensors at any time t, obtained by taking the center of mass feature point of the human body as the origin.

[0036] R t =(C 实测t-1 C 实测t-2 C 标准t-i , ...C 标准t-N ) T C 实测t-i Let be the coordinate matrix measured by the i-th attitude sensor; R t ’ =(C 实测t-1 -P st C 实测t-2 -P st C 标准t-i -P st , ...C 标准t-N -P st ) T R t To R t ’ Essentially, it is the relative coordinate matrix of the attitude sensor corresponding to N measured actions at any time t, obtained by taking the human body's center of mass feature point as the origin.

[0037] S400, solve for R0 ’’ ~R T ’’ ; The standard point cloud registration algorithm (singular value decomposition) is used to calculate and apply the optimal rotation matrix to eliminate rotation errors, thereby obtaining the real-time registration matrix R at any time t that is spatially aligned with the standard action pose. t ’’ ; R t ’’ The solution method includes the following sub-steps: S401, via M t ’ and R t ’ Solve for the cross-covariance matrix H t ; H t =M t ’T ·R t ’ ; M t’T It is M t ’ The transpose of the matrix; S402, for H t Perform SVD decomposition (for the SVD decomposition algorithm, please refer to: https: / / blog.csdn.net / qq_61636702 / article / details / 140458126): H t =V t ·S t ·W t T V t It is a 3×3 orthogonal matrix, S t It is a 3×3 diagonal matrix, W t It is a 3×3 orthogonal matrix; S403, Solving for the optimal rotation matrix U t : S4031, U t =W t ·V t T ; S4032, to ensure the calculated matrix U t It is a pure rotation matrix (determinant is +1) rather than a reflection matrix (determinant is -1), so a verification is needed. Calculate U. t The determinant det(U t If det(U) t )=+1, then U t This is the optimal rotation matrix we are looking for. If det(U) t If )=-1, it indicates that a reflection solution has appeared, usually due to data noise or symmetry.

[0038] That is, if det(U t )>0, U t No corrections will be made; If det(U) t If W < 0, then W t Invert the third column (i.e., multiply by -1) and then combine it with V. t T The product of U t The result; det(U t ) represents U t The determinant result; S404, R t ’’ =R t ’ ·U t ; S500 solves the direction vector of each bone (e.g., left upper arm, left lower arm, right upper arm, right lower arm, left thigh, left calf, right thigh, right calf) representing the standard movement and the measured movement, as well as the joint angle between two adjacent bones.

[0039] For example: M t R t The first and second rows of data contain the sensor coordinates of the left upper arm and left elbow joint. Therefore, for the left upper arm, its direction vector can be M. t ’ R t ’’ The data from the second row to the first row is used as the direction vector.

[0040] From this, the direction vectors V of the first and second skeletons representing the standard movement can be obtained. std1 V std2 and joint angle α std ; α std =arccos((V std1 ·V std2 ) / (|V std1 |×|V std2 |)) From this, the direction vectors V of the first and second skeletons representing the measured action can be obtained. real1 V real2 and joint angle α real ; α real =arccos((V real1 ·V real2 ) / (|V real1 |×|V real2 |)) S600, Logical Comparison: If |α std -α real |≤δ 阈值 β1≤β 1阈值 β2≤β 2阈值 The action meets the requirements; Otherwise, the actions do not match; Wherein, the angle β1 between the first skeleton of the measured movement and the first skeleton of the standard movement is β1 = acrcos((V real1 ·V std1 ) / (|V real1 |×|V std1 |)) Among them, the angle β between the second skeleton of the measured movement and the second skeleton of the standard movement. 2= =acrcos((Vreal2 ·V std2 ) / (|V real2 |×|V std2 |)) Where, δ 阈值 This represents the threshold for joint angle motion matching between the first and second bones, in degrees. β 1阈值 This represents the threshold angle for the change in the direction of movement of the first bone, in degrees. β 2阈值 This represents the threshold angle for the change in the direction of movement of the second bone, expressed in degrees.

[0041] <Example 4: A Method for Assessing Muscle Status During Rehabilitation Exercise> A method for assessing muscle status during rehabilitation exercises, employing two modes during imaging: Step A, First Operating Mode: Global Low-Precision Physiological Activity Monitoring and Linear Back-Projection (LBP) Reconstruction; The method of this invention, by default, continuously operates in a global low-precision physiological activity monitoring mode, i.e., the first operating mode. The core objective of this mode is not to perform precise quantification of physiological parameters, but rather to conduct a rapid and continuous survey of the activity status of all major muscle groups covered by sensor components throughout the user's body while ensuring extremely high temporal resolution. The system employs the Linear Back-Projection (LBP) algorithm, which simplifies the complex nonlinear EIT inverse problem into a direct linear calculation.

[0042] △σ LBP = J T ·△V; Where ΔV is the boundary voltage difference matrix measured by the impedance sensor assembly placed in any muscle of the subject; J is the conductivity sensitivity matrix (Jacobian matrix), which describes the effect of small changes in conductivity of each cell in the imaging region on the boundary voltage measurement; Δσ LBP The final output is not a traditional EIT image used for diagnosis, but a qualitative visualization called a "Physiological Activity Heatmap," which is a matrix of changes in unit conductivity. This heatmap is rendered in real-time and overlaid on a three-dimensional biomechanical model driven by visual and IMU data. The heatmap visually represents the degree of impedance change of various muscle groups relative to their resting baseline state, using different colors or brightness levels. For example, warmer colors or higher brightness indicate more intense muscle contraction or physiological activity in that area. Users or system operators can easily grasp the real-time, dynamic activation overview of all muscles in the body through this heatmap.

[0043] Dynamic Region of Interest (ROI) identification and high-precision imaging triggering; the system presets a trigger threshold Δσ. t When |Δσ| exceeds this threshold, the system determines that a significant physiological event of analytical value has occurred. The determination condition can be expressed as: |Δσ|>Δσ t .

[0044] Once a region that meets the triggering conditions is detected, the system automatically defines that region as a Dynamic Region of Interest (ROI) and records its location, extent, and trigger time. ROI identification immediately generates an internal software interrupt or event signal. This signal serves as an explicit instruction, triggering the system to switch from a first operating mode of global low-precision monitoring to a second operating mode of targeted high-precision analysis of that specific ROI.

[0045] In other words, the muscle is normally imaged using the first operating mode to save computational resources. When a Region of Interest (ROI) is encountered, the system switches to the second operating mode, which enables high-precision imaging. For example, when the upper arm muscle is an ROI, step B is executed for the upper arm muscle.

[0046] Step B: Targeted high-precision EIT reconstruction and analysis; Reconstruction algorithms for conductivity distribution images typically use the Gauss-Newton method (see: Wang Y, Feng H, Sun B, et al. Plantar Planar Dual Array Electrical Impedance Tomography (PPDA-EIT) Method for Early Diabetic Foot Ulcers Detection[J]. IEEE SensorsJournal, 2024.) or other methods to calculate the time-difference relative conductivity matrix and then perform imaging.

[0047] Upon receiving the trigger signal from step A, the computer-controlled impedance analyzer injects current with frequencies ranging from f1 (low frequency) to f2 (high frequency) into the human body from a pair of electrodes. Simultaneously, the impedance data of the remaining adjacent electrode pairs are measured, and the impedance analyzer acquires impedance spectrum data once. Then, the phase angle distribution image is reconstructed using a phase angle image reconstruction algorithm (see reference: Sun B, Sejati PA, Shirai T, et al. Long-term phase angle muscle imaging under electrical muscle stimulation (EMS) by phase angle electrical impedancetomography[J]. Physiological Measurement, 2024, 45(7): 075010) to characterize the degree of muscle activation.

[0048] After imaging, the state of the muscle region is calculated; Muscle mass index: Calculate the average relative electrical conductivity value Δσ of muscle. mean This is calculated as the average value of the time-difference relative conductivity matrix. Conductivity primarily reflects the content of conductive media (such as water and electrolytes) in tissues. The water and electrolyte content in muscle tissue is much higher than that in fat, therefore, conductivity is positively correlated with lean body mass (i.e., muscle mass).

[0049] Sarcoplasmic index: Calculate the spatial mean phase angle Δφ of muscle. mean (See reference: Sun B, Sejati PA, Shirai T, et al. Long-term phase angle muscle imaging under electrical musclestimulation (EMS) by phase angle electrical impedance tomography[J]. Physiological Measurement, 2024, 45(7): 075010). Healthy muscle cells have intact cell membranes, which results in higher body electrical impedance values, leading to a higher phase angle. Therefore, phase angle is often considered an effective indicator for assessing sarcoplasm (i.e., the health and quality of muscle cells).

[0050] When Δσ mean =0, Δφ mean =0 indicates that the muscle is inactive.

[0051] When Δσ mean ≠0 or △φ mean≠0 indicates that the muscle is in an activated state: According to △σ mean To assess muscle mass: (1) Δσ mean ≤a 阈值 This indicates that the muscle content is low; (2) a 阈值 <△σ mean ≤b 阈值 This indicates that the muscle mass is within the normal range; (3)b 阈值 <△σ mean This indicates a high muscle mass; among them, a 阈值 b 阈值 It is a threshold; According to △φ mean To assess skin texture: (1) △φ mean ≤u 阈值 This indicates that the health and integrity (sarcoplasm) of muscle cells have been affected; (2) u 阈值 <△φ mean ≤v 阈值 This indicates that the muscle cells are healthy; (3) v 阈值 <△φ mean This indicates that muscle cells have been enhanced; among them, u 阈值 v 阈值 It is a threshold.

[0052] For rehabilitation exercises, the following situations generally occur: △σ mean ≤a 阈值 , △φ mean <u 阈值 This indicates poor muscle quality and cell function, requiring enhanced intervention. △σ mean ≤a 阈值 u 阈值 <△φ mean This occurs in the early stages of rehabilitation or training. The patient's total muscle mass (reflected in relative conductivity) remains low due to prolonged inactivity or disease, but due to active rehabilitation or training, the health and integrity of muscle cells (reflected in phase angle) begin to improve significantly. This indicates that the muscles are actively responding to exercise, a positive sign.

[0053] a 阈值 <△σ mean , △φ mean ≤u 阈值 This indicates that the total muscle mass (muscle mass) has not decreased significantly and remains within the normal range. This is because the patient's lean body mass remains at a relatively healthy level. However, the health and integrity of muscle cells (sarcoplasm) have been affected. This is due to factors such as impaired muscle cell membrane function, cell atrophy, or tissue edema.

[0054] a 阈值 <△σ mean ≤b 阈值 u 阈值 <△φ mean This indicates that muscle mass is normal and muscle cells are in a healthy state, such as in long-term fitness enthusiasts.

[0055] It should be noted that the probability of the two cases, "△σmean≠0,△φmean=0" or "△σmean=0,△φmean≠0", is very small, because the fluctuation of muscle mass and muscle quality will cause these two values ​​to be non-zero, so they do not need to be discussed.

[0056] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made to the technical solutions of the present invention by those skilled in the art without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A method for assessing rehabilitation exercises, characterized in that, It includes the following steps: In the first stage, the standard assessment method for rehabilitation exercise movements is used to evaluate whether the movements are standard. If the movements meet the requirements, the second stage is carried out; otherwise, the second stage is not carried out. The standard assessment method for rehabilitation exercise movements is as follows: the movement is assessed by comparing the direction of any two adjacent bones representing the pre-set standard movement and the joint angle between the two bones with the corresponding direction of the two bones representing the actual measured movement and the corresponding joint angle. In the second stage, the muscle status is assessed using the rehabilitation exercise muscle status assessment method. The method for assessing muscle condition during rehabilitation exercises includes the following steps: Step a: Acquire an image of the muscle and extract its contour line, using the area within the contour line as the EIT simulation model area; read the boundary voltage difference matrix ΔV of the muscle, and based on Δσ... LBP =J T · Solve for the element conductivity change matrix △σ using △V LBP Where J is the conductivity sensitivity matrix, and ΔV is the time-difference boundary voltage difference matrix of the muscle to be tested. If matrix Δσ LBP The absolute value of any element in the array is greater than the preset threshold Δσ. t If the muscle is found to have a dynamic region of interest, then step b is executed; otherwise, step b is not executed. Step b, perform targeted high-precision EIT reconstruction and analysis of the muscle: solve the time-difference relative conductivity matrix of the muscle by using the boundary voltage difference matrix ΔV, and output the reconstructed conductivity distribution image; Step c: Based on the time difference relative conductivity matrix obtained in step b, calculate the average value Δσ of the time difference relative conductivity matrix. mean According to △σ mean To determine the muscle mass status: when Δσ mean =0 indicates that the muscle is in an inactive state; when Δσ = 0, it indicates that the muscle is inactive; mean ≠0 indicates that the muscle is in an activated state: (1) △σ mean ≤a 阈值 This indicates a relatively small total muscle mass; (2) a 阈值 <△σ mean ≤b 阈值 This indicates that the muscle mass is within the normal range; (3)b 阈值 <△σ mean This indicates a high muscle mass; among them, a 阈值 b 阈值 It is a threshold.

2. The rehabilitation exercise assessment method according to claim 1, characterized in that, The assessment methods for rehabilitation exercise standards also include: The method for evaluating any two adjacent bones at any time t: the first bone, the second bone, and the joint angle between them includes the following steps: S100, obtain M and P in the same coordinate system; , , Where M is the coordinate matrix of the three attitude sensors representing the standard action; (x) 标准-1 y 标准-1 , z 标准-1 ), (x 标准-2 y 标准-2 , z 标准-2 ), (x 标准-3 y 标准-3 , z 标准-3 () represent the three-dimensional coordinates of the posture sensors of the first bone, joint, and second bone, respectively, representing the standard movement; P is the coordinate matrix of the human body's center of mass characteristic points during standard movements (x 质心-标准 y 质心-标准， z 质心-标准 () represents the three-dimensional coordinates of the human body's center of mass feature point during standard movements; S200, obtain R and P in the same coordinate system S ; , , Where R is the coordinate matrix of the three attitude sensors representing the measured action; (x 实测-1 y 实测-1 , z 实测-1 ), (x 实测-2 y 实测-2 , z 实测-2 ), (x 实测-3 y 实测-3 , z 实测-3 ) represent the three-dimensional coordinates of the posture sensors at the first bone, joint, and second bone, respectively, representing the actual measured movement; Among them, P S The coordinate matrix representing the human body's center of mass during the measured action; (x 质心-实测 y 质心-实测， z 质心-实测 () represents the three-dimensional coordinates of the human body's center of mass feature point during the measured action; S300, solve for the relative coordinate matrix M' of the three attitude sensors when representing the standard action, and the relative coordinate matrix R' of the three attitude sensors when representing the measured action; , , S400, solve for the real-time registration matrix R'', which is solved using steps S401 to S404; S401, Solve for the cross-covariance matrix H using M' and R': H=M' T ·R', where M' T It is the transpose of M'; S402, Perform SVD decomposition on H: H = V·S·W T V is a 3×3 orthogonal matrix, S is a 3×3 diagonal matrix, and W is a 3×3 orthogonal matrix; S403, Solve for the optimal rotation matrix U: S4031, U=W·V T V T It is the transpose of V; S4032, if det(U)>0, U is not modified; if det(U)<0, the third column of W is inverted and then ANDed with V. T The product of these is the result of U; S404, R''=R'·U; R'' is expressed as: , Where, x 实测-1 ''、x 实测-2 ''、x 实测-3 ''、y 实测-1 ''、y 实测-2 ''、y 实测-3 ''、z 实测-1 ''、z 实测-2 ''、z 实测-3 '' is an element of R'; S500, solve for the direction vectors V of the first and second bones when representing the standard movement. std1 V std2 and joint angle α std ; V std1 =(x 标准-1 -x 标准-2 ,y 标准-1 -y 标准-2 ,z 标准-1 -z 标准-2 ); V std2 =(x 标准-3 -x 标准-2 ,y 标准-3 -y 标准-2 ,z 标准-3 -z 标准-2 ); α std =arccos((V std1 ·V std2 ) / (|V std1 |×|V std2 |)) S600, solve for the direction vectors V of the first and second skeletons when representing the measured motion. real1 V real2 and joint angle α real ; V real1 =(x 实测-1 ''-x 实测-2 '',y 实测-1 ''-y 实测-2 '',z 实测-1 ''-z 实测-2 ''); V real2 =(x 实测-3 ''-x 实测-2 '',y 实测-3 ''-y 实测-2 '',z 实测-3 ''-z 实测-2 ''); α real =arccos((V real1 ·V real2 ) / (|V real1 |×|V real2 |)) S700, Logical Comparison: If |α std -α real |≤δ 阈值 β1≤β 1阈值 β2≤β 2阈值 The action meets the requirements; Otherwise, the actions do not match; Wherein, the angle β1 between the first skeleton of the measured movement and the first skeleton of the standard movement is β1 = acrcos((V real1 ·V std1 ) / |V real1 |×|V std1 |); Among them, the angle β between the second skeleton of the measured movement and the second skeleton of the standard movement. 2= =acrcos((V real2 ·V std2 ) / |V real2 |×|V std2 |); Where, δ 阈值 This represents the threshold for joint angle motion matching between the first and second bones, in degrees. β 1阈值 This represents the threshold angle for the change in the direction of movement of the first bone, in degrees. β 2阈值 This represents the threshold angle for the change in the direction of movement of the second bone, expressed in degrees.

3. The rehabilitation exercise assessment method according to claim 1, characterized in that, The direction vectors of any two adjacent bones and the joint between the two bones are: the direction vectors of the left upper arm and left forearm and the angle of the left elbow joint. And / or, the direction vectors of the right upper arm and right forearm, and the angle of the right elbow joint; And / or, the direction vectors of the left thigh and left calf, and the angle of the left knee joint; And / or, the direction vectors of the right thigh and right calf, and the angle of the right knee joint.

Images