Gluteus maximus erector spinae pectoralis major multi-modal assessment and health and wellness matching system

Abstract

The application discloses a gluteus maximus erector spinae pectoralis major multi-modal evaluation and health care matching system and belongs to the technical field of intelligent health care. The system comprises a multi-modal data acquisition unit, an AI feature extraction and state evaluation unit, a longitudinal comparison storage unit and a health care scheme intelligent matching output unit. An adaptive high-frequency ultrasonic probe, a multi-frequency multi-electrode body composition module and an RGB-D depth camera are used to directionally collect four-dimensional physical parameters of the muscle thickness value, echo gray mean value, fat infiltration proportion and body posture offset of the gluteus maximus, erector spinae and pectoralis major. An AI image segmentation network automatically completes muscle region recognition and parameter extraction, a multi-parameter fusion evaluation model outputs four state grades, and a longitudinal comparison submodule measures muscle group thickness change rate and generates trend estimation. The evaluation result is automatically matched with daily nutritional meal suggestions and home activity guidance schemes to form an integrated health care service closed loop of 'evaluation-suggestion-follow-up'. The application realizes automatic extraction of physical parameters and state evaluation of three core muscle groups and is suitable for community health care centers, old-age care institutions, family health care and other multi-scene popularization and use.

Description

Technical Field

[0001] To address the shortcomings of existing technologies, such as the inability to quantify the physical parameters of local muscle groups like the gluteus maximus, erector spinae, and pectoralis major independently, reliance on manual operation for assessment, and the disconnect between detection and rehabilitation recommendations, this invention provides an intelligent matching system for assessing the condition of the gluteus maximus, erector spinae, and pectoralis major based on multimodal sensor fusion and matching rehabilitation plans. This system uses an adaptive high-frequency ultrasound probe, a multi-frequency multi-electrode body composition module, and an RGB-D depth vision camera to directionally acquire multi-dimensional physical parameters of the three core muscle groups. An AI algorithm automatically performs muscle region segmentation and feature extraction, outputting quantitative parameters such as muscle thickness, mean echo grayscale, fat infiltration percentage, and body posture deviation. Based on these parameters, it automatically matches daily nutritional dietary recommendations and home exercise guidance plans, constructing an integrated rehabilitation service loop of "assessment-recommendation-follow-up." Background Technology

[0002] In the context of an aging population, declining muscle mass is a significant factor affecting the mobility and quality of life of middle-aged and elderly individuals. Regular assessment and health management of the three core muscle groups—the gluteus maximus, erector spinae, and pectoralis major—have significant social value. However, current muscle assessment and rehabilitation techniques have the following obvious shortcomings: Mismatched detection targets: Conventional body composition equipment only outputs the total amount of skeletal muscle in the limbs as a whole, and cannot separately quantify physical parameters such as muscle thickness, echo texture, and fat infiltration of local muscle groups such as the gluteus maximus, erector spinae, and pectoralis major, which cannot meet the needs of fine-grained status assessment for specific muscle groups. Assessment relies on manual operation: Traditional musculoskeletal ultrasound relies on physicians to manually interpret images. The assessment results are greatly affected by the differences in the operator's experience, making it difficult to achieve standardized and repeatable automatic quantitative extraction of parameters such as muscle thickness, grayscale texture, and fat infiltration ratio. Lack of longitudinal tracking capability: Existing assessment schemes are mostly single-section assessments, lacking cross-cycle longitudinal data comparison, state change rate measurement and trend prediction mechanisms, and cannot provide continuous state change tracking for individuals; Testing and health advice are disconnected: Current testing equipment only issues a data report after completing the assessment, and cannot automatically match daily nutritional diet suggestions and home exercise guidance programs based on the test results, thus failing to form a closed-loop health service process of "assessment-advice-follow-up". In existing publicly available technologies, muscle assessment targets are mostly focused on the lower limb muscles such as the gastrocnemius and quadriceps. There is no system specifically designed for the three core muscle groups of the gluteus maximus, erector spinae, and pectoralis major to automatically extract multi-dimensional physical parameters and assess their status. Furthermore, there is no health and wellness equipment that can automatically match and link the results of muscle group status assessment with daily nutritional dietary recommendations and home exercise guidance programs. Summary of the Invention

[0003] Purpose of the invention To address the shortcomings of existing technologies, such as the inability to quantify the physical parameters of local muscle groups like the gluteus maximus, erector spinae, and pectoralis major independently, reliance on manual operation for assessment, and the disconnect between detection and rehabilitation recommendations, this invention provides a multimodal assessment and rehabilitation matching system for the gluteus maximus, erector spinae, and pectoralis major muscles. This system uses an adaptive high-frequency ultrasound probe, a multi-frequency multi-electrode body component module, and an RGB-D depth vision camera to directionally acquire multidimensional physical parameters of the three core muscle groups. An AI algorithm automatically performs muscle region segmentation and feature extraction, outputting quantitative parameters such as muscle thickness, mean echo grayscale value, fat infiltration percentage, and body posture deviation. Based on these parameters, it automatically matches daily nutritional dietary recommendations and home exercise guidance programs, constructing an integrated rehabilitation service loop of "assessment-recommendation-follow-up." Technical solution A multimodal assessment and rehabilitation matching system for the gluteus maximus, erector spinae, and pectoralis major muscles includes a multimodal data acquisition unit, an AI feature extraction and status assessment unit, a longitudinal comparison and storage unit, and a rehabilitation plan intelligent matching output unit; the entire system is integrated within a rehabilitation workstation. Multimodal data acquisition unit: 1) Adaptive high-frequency musculoskeletal ultrasound probe: used to acquire muscle thickness values, ultrasound echo grayscale texture features, and intramuscular fat infiltration ratio of the gluteus maximus, erector spinae, and pectoralis major muscles. It has a wideband adaptive adjustment function, and the system spatial resolution reaches the sub-millimeter level; 2) Multi-frequency multi-electrode human composition detection module: acquires body composition parameters such as total skeletal muscle mass, bioimpedance value, and phase angle as an auxiliary reference baseline for muscle group status assessment; 3) RGB-D depth vision camera: acquires human body contour, thoracic morphology, and pelvic tilt angle to complete the quantitative acquisition of body posture deviation; 4) Basic vital signs input module: inputs basic information such as the subject's age, height, weight, and average daily sitting time as auxiliary input features for the model. AI Feature Extraction and State Assessment Unit 1) Image segmentation network module: A deep learning image segmentation network is used to perform semantic segmentation on ultrasound images. It automatically selects the regions of interest for the gluteus maximus, erector spinae, and pectoralis major muscles, removes interference signals such as subcutaneous fat and bone echoes, and extracts four core physical parameters: muscle thickness (unit: mm), average gray value of muscle echo (0-255 gray value range), intramuscular fat replacement rate (%), and body posture offset (angle value, °). 2) Multi-parameter fusion assessment model: Based on the extracted multi-dimensional physical parameters, a machine learning fusion model outputs muscle group state assessment results, specifically divided into four state levels: Grade 0: Parameters such as muscle thickness and grayscale are within the normal fluctuation range; Grade I: No significant decrease in muscle thickness, but a slight increase in echogenicity (suggesting changes in tissue elasticity). Grade II: A quantifiable slight decrease in muscle thickness and a slight increase in the percentage of fat infiltration; Grade III: Muscle thickness is significantly reduced, and the proportion of fat infiltration is significantly increased; 3) Vertical comparison submodule: Stores all detection data in the cloud, automatically calculates the annual rate of change of the thickness of the three major muscle groups, generates a state change trend curve, and outputs the estimated data of the muscle group thickness change trend in the next 1-3 years. The longitudinal comparison storage unit establishes an individual's electronic health record, saving the original images, feature data, status level, and trend curves of each screening, and supports cross-cycle review and comparison, data export, and long-term follow-up management. The intelligent matching and output unit for health and wellness solutions receives AI status assessment results and automatically matches and outputs non-medical health and wellness recommendations based on quantitative parameters such as muscle thickness and fat infiltration percentage. 1) Push daily nutritional dietary suggestions to the workstation touch interface, including daily intake reference suggestions and consumption cycle reminders for foods such as high-quality protein and calcium; 2) Provide home activity guidance programs – including illustrated guides for daily stretching and low-intensity activities targeting the three core muscle groups; 3) Generate and update the long-term follow-up plan, indicating the next check-up time and key points to focus on. Furthermore, this system is an integrated hardware and software device, not just a simple algorithm program. The ultrasound probe, body composition module, and depth camera serve as hardware acquisition carriers, the AI ​​algorithm serves as the built-in data processing program, and the touch interactive display screen serves as the solution output terminal, forming a complete health and wellness assessment and advice device. This invention also provides a multimodal assessment and rehabilitation matching system for the gluteus maximus, erector spinae, and pectoralis major muscles, comprising the following steps: S1: Subjects enter basic vital sign data and complete ultrasound image acquisition, multi-frequency bioimpedance detection, and RGB-D body posture scanning for the three core muscle groups. S2: AI algorithm automatically segments the target muscle region and extracts multi-dimensional physical parameters such as muscle thickness, average echo grayscale, fat infiltration ratio, and body posture offset. S3: The multi-parameter fusion evaluation model outputs 0-III level state evaluation results, calculates the annual rate of change of muscle group thickness, and generates a state change trend prediction report. S4: The system automatically matches daily nutritional dietary recommendations and home activity guidance plans based on quantitative parameters such as muscle thickness and fat infiltration ratio, and completes integrated recommendation push; S5: Establish follow-up records, automatically retrieve historical data for subsequent re-examinations, and dynamically update the status assessment level and trend. Beneficial effects Innovative detection target: Abandoning the general approach to detecting limb muscle groups, it focuses on three core muscle groups: gluteus maximus, erector spinae, and pectoralis major, filling the technological gap in the automated extraction and status assessment of physical parameters such as muscle thickness, echo texture, and fat infiltration for specific local muscle groups. Automated assessment innovation: The use of deep learning image segmentation network to achieve automatic identification and parameter extraction of muscle regions solves the technical problems of traditional ultrasound assessment relying on manual interpretation and non-reproducible results. It realizes the standardized and repeatable quantitative extraction of parameters such as muscle thickness, grayscale, and fat infiltration. Longitudinal tracking innovation: Establish a cross-cycle longitudinal data comparison mechanism to automatically calculate the rate of change in muscle group thickness and generate trend predictions, providing individuals with the ability to continuously track changes in their condition; Multimodal sensing fusion innovation: It organically integrates three sensing modalities, namely high-frequency ultrasound, multi-frequency bioimpedance, and depth vision, into a single workstation to achieve synchronous acquisition of multi-dimensional physical parameters from muscle microstructure to macroscopic body posture. Assessment-Recommendation Closed-Loop Innovation: Test results directly drive the health and wellness workstation to output daily nutritional dietary recommendations and home activity guidance programs, solving the industry pain point of the traditional disconnect between testing and health and wellness recommendations; Clear Product Positioning: This invention is clearly positioned as a non-medical device for health assessment and advice. The system outputs quantitative values ​​of physical parameters such as muscle thickness, grayscale, and fat infiltration, as well as corresponding daily nutrition and activity recommendations. It does not output any disease diagnosis conclusions and is suitable for promotion and use in multiple scenarios such as community health centers, elderly care institutions, and family health care. Attached Figure Description Figure 1 This is a block diagram of the overall system architecture of the present invention; Figure 2 A closed-loop workflow diagram for assessment and recommendations; Figure 3 A schematic diagram of region of interest segmentation for the target muscle group in ultrasound imaging. Figure 4 This is a schematic diagram of the output interface for the status assessment level. Detailed Implementation The present invention will be further described in detail below with reference to the embodiments. Example 1 A multimodal assessment and rehabilitation matching system for the gluteus maximus, erector spinae, and pectoralis major muscles is provided. The hardware components are assembled in a rehabilitation workstation host: a wideband adaptive high-frequency linear ultrasound probe (frequency range 7-15MHz, adaptively adjustable); a multi-frequency multi-electrode body component module (frequency points: 5kHz, 50kHz, 250kHz, 500kHz); an RGB-D depth camera (depth resolution 640×480@30fps, angle measurement accuracy ≤2°); the industrial control host has a built-in AI image segmentation network (based on U-Net architecture, with ResNet-34 as the backbone network) and a multi-parameter fusion evaluation model (based on random forest algorithm, fusing four-dimensional features of muscle thickness, grayscale mean, fat infiltration rate, and body posture offset); and the workstation is equipped with a touch interactive display screen. AI Model Training and Validation The image segmentation network was trained using a dataset containing at least 3000 ultrasound images of the gluteus maximus, erector spinae, and pectoralis major muscles. All images were manually annotated by at least three radiologists with over five years of experience in musculoskeletal ultrasound, serving as the gold standard. The segmentation network achieved a Dice coefficient ≥ 0.85 on the validation set. The multi-parameter fusion evaluation model was calibrated using MRI-measured muscle cross-sectional area and fat fraction as reference standards. The accuracy of each physical parameter extraction was verified, with an intraclass correlation coefficient (ICC) ≥ 0.80 between the extracted parameters and the MRI measurements. Subject testing procedure: 1) Subjects entered basic information such as age, height, weight, and daily sedentary time on the touch screen; 2) The operator used an adaptive high-frequency ultrasound probe to sequentially acquire ultrasound images of the gluteus maximus, erector spinae, and pectoralis major muscles—the gluteus maximus was acquired in the prone position using surface landmarks of the buttocks, the erector spinae in the prone position using surface landmarks of the lumbar region, and the pectoralis major in the supine position using surface landmarks of the chest—an AI image segmentation network automatically measured the muscle thickness of each muscle group (accurate to 0.1 mm) and extracted the mean echo grayscale value (0-255 grayscale value) and the percentage of fat infiltration (%). 3) The subject stands with both feet on the multi-electrode component platform, and completes the acquisition of multi-frequency impedance and phase angle data in 15 seconds; 4) RGB-D camera scans and acquires pelvic tilt angle and thoracic posture data, and calculates posture offset (°); 5) The AI ​​feature extraction and state assessment unit integrates all feature data and outputs the state assessment level: 0 / Ⅰ / Ⅱ / Ⅲ. It calculates the annual change rate of muscle group thickness (mm / year) and outputs the predicted trend of muscle group thickness change in the next 1-3 years. 6) Based on quantitative parameters such as muscle thickness and fat infiltration percentage, the system automatically matches daily nutritional dietary recommendations (such as daily reference intake of high-quality protein and daily pairing recommendations of calcium-rich foods) and home activity guidance programs (such as graphic guidance on gluteus maximus activation training, erector spinae stretching training, and pectoralis major flexibility training), which are pushed and displayed via touch screen and can be exported and printed. 7) All assessment reports and health care recommendations are saved to your personal electronic file, and historical data will be automatically retrieved for comparison during the next follow-up visit. Example 2 As a workstation expansion module, this system can be independently configured as a desktop health and wellness assessment terminal. Without the configuration of physiotherapy hardware, it can only perform muscle group status assessment, health and wellness plan push and file management functions, and is suitable for use in community health and wellness centers and elderly care institutions.

Claims

1. A multimodal assessment and rehabilitation matching system for the gluteus maximus, erector spinae, and pectoralis major muscles, characterized in that, include: Multimodal data acquisition unit, AI feature extraction and status assessment unit, longitudinal comparison and storage unit, and intelligent matching and output unit for health and wellness solutions; The multimodal data acquisition unit includes an adaptive high-frequency ultrasound probe, a multi-frequency multi-electrode body composition module, an RGB-D depth vision camera, and a vital sign recording module. The adaptive high-frequency ultrasound probe is used to acquire muscle thickness values, ultrasound echo grayscale texture features, and intramuscular fat infiltration percentages of the gluteus maximus, erector spinae, and pectoralis major muscles. The multi-frequency multi-electrode body composition module is used to acquire the total amount of skeletal muscle and bioimpedance values. The RGB-D depth vision camera is used to acquire the human body contour and pelvic tilt angle and output the body posture offset. The AI ​​feature extraction and state assessment unit includes an image segmentation network module, a multi-parameter fusion assessment model, and a longitudinal comparison submodule. The image segmentation network module uses a deep learning image segmentation network to perform semantic segmentation on ultrasound images, automatically selecting regions of interest for the gluteus maximus, erector spinae, and pectoralis major muscles, and extracting four-dimensional physical parameters: muscle thickness, mean echo grayscale, fat infiltration percentage, and body posture offset. The longitudinal comparison submodule uses historical change rates as prior weights input into the multi-parameter fusion assessment model, dynamically adjusting the assessment model's judgment threshold. The body posture offset extracted by the image segmentation network module is used for spatial coordinate correction of the segmented regions in the ultrasound image. The multi-parameter fusion assessment model outputs four state levels based on the extracted four-dimensional physical parameters: level 0, level I, level II, and level III. The longitudinal comparison submodule calculates the annual change rate of thickness for each muscle group, generating state change trend prediction data. The intelligent matching output unit of the health and wellness plan automatically matches daily nutritional dietary recommendations and home activity guidance plans based on quantitative parameters such as muscle thickness and fat infiltration percentage, and outputs them through the workstation touch interface.

2. The system according to claim 1, characterized in that, The system is integrated into a health and wellness workstation, with hardware data acquisition units, AI processing units, and touch-screen interactive output units assembled in one unit, forming a health and wellness assessment and advice device that combines software and hardware.

3. The system according to claim 1, characterized in that, The image segmentation network module adopts a deep learning segmentation network based on the U-Net architecture, with ResNet as the backbone network. The muscle thickness value is accurate to 0.1mm, the mean echo gray value is a quantitative value in the gray value range of 0-255, the fat infiltration ratio is expressed as a percentage (%), and the body position offset is expressed as an angle value (°).

4. The system according to claim 1, characterized in that, The adaptive high-frequency ultrasound probe has a wide-band adaptive adjustment function with a frequency range of 7-15MHz, and the system spatial resolution reaches the sub-millimeter level.

5. The system according to claim 1, characterized in that, The daily nutritional dietary recommendations output by the intelligent matching output unit of the health and wellness program include daily intake reference suggestions and dosage cycle reminders for high-quality protein and calcium foods; the home activity guidance program includes illustrated guides for daily stretching and low-intensity activities for the gluteus maximus, erector spinae, and pectoralis major muscles.

6. A multimodal assessment and rehabilitation matching method for the gluteus maximus, erector spinae, and pectoralis major muscles, characterized in that... Includes the following steps: S1: Enter basic vital sign data, and sequentially complete ultrasound acquisition of the three core muscle groups, multi-frequency impedance detection, and RGB-D body posture scanning. S2: AI algorithm segmentes muscle regions and extracts four-dimensional physical parameters: muscle thickness, mean echo grayscale, fat infiltration percentage, and body posture offset. S3: The multi-parameter fusion evaluation model outputs a 0-III level state evaluation grade and calculates the annual change rate of thickness of each muscle group. S4: Based on quantitative parameters such as muscle thickness and fat infiltration percentage, it automatically matches daily nutritional dietary recommendations and home activity guidance programs. S5: Establish follow-up records and review and update the status assessment results and trends.

Images