Multidimensional motion state evaluation method and system assisted based on voice biomarker
By collecting and fusing the features of speech and heart rate signals in real time, and using a multimodal fusion model to generate multidimensional motion state assessment results, the problem of lack of deep heterogeneous fusion of speech and heart rate signals in existing technologies is solved, and more accurate and personalized motion state assessment is achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- XIANGYA HOSPITAL CENT SOUTH UNIV
- Filing Date
- 2026-04-09
- Publication Date
- 2026-07-10
AI Technical Summary
In existing technologies, there is a lack of deep heterogeneous fusion between voice and heart rate signals, making it difficult to extract stable physiological features under motion interference, resulting in one-sided and delayed motion state assessment results.
By acquiring speech and heart rate signals in real time, separating the human voice segment from the silent segment, extracting time-domain breathing features and frequency-domain stability features, removing motion artifacts from the heart rate signal, and using a multimodal fusion model to perform dual-path feature extraction and multi-head attention fusion, a multidimensional motion state assessment result is generated.
It improves the comprehensiveness, real-time performance, anti-interference capability, and personalized guidance accuracy of motion status assessment, provides personalized motion intensity adjustment instructions, and enhances the accuracy and reliability of the assessment.
Smart Images

Figure CN122350690A_ABST