Examining the Relationship Between Hand Gesture Recognition Accuracy and Data Quality Indices for EMG and pFMG Signals: A Correlation Analysis

Accurate recognition of hand gestures is crucial especially for in prosthetic control and human-machine interactions, and other applications. This study explores the correlation between data quality indices including Signal-to-Noise Ratio (SNR) and Signal-to-Motion Artifact Ratio (SMR), and hand gesture recognition accuracy, utilizing Surface Electromyography (sEMG) and pressure-based Force Myography (pFMG) signals. Through this investigation, we demonstrate that prioritizing SNR in sEMG signals and SMR in FMG signals when positioning the sensors has the potential to improve hand gesture recognition performance, as our analysis showed that SNR has a higher correlation with hand gesture accuracy with sEMG signals while SMR has a higher correlation with hand gesture accuracy with pFMG signals. The findings highlight the critical role that comprehensive data quality assessments play in improving the accuracy of the hand gesture recognition and the reliability of prosthetic control systems, for example. By enhancing the signal quality for both sEMG and pFMG, this study provides new insights into sensor positioning and channel selection for the development of more efficient and precise gesture recognition systems for assistive devices. It emphasizes the necessity of utilizing comprehensive data quality assessment and lays the foundation for future enhancements in gesture recognition accuracy by addressing signal contamination or quality challenges.