A Signal Elevation Approach for Enhanced Detection of Onset and Offset of Muscle Activation Segments in Surface Electromyography

This paper presents a comprehensive study on the development and validation of a surface Electromyography (sEMG) onset-offset detection algorithm. sEMG signals have proven to be valuable in various applications, particularly in gesture recognition. Accurate detection of the onset and offset points in sEMG signals is crucial in the context of precise gesture recognition. In this work, we introduce a novel signal elevation algorithm designed to enhance the gesture region within sEMG data, which significantly simplifies the onset-offset detection process. The proposed algorithm is thoroughly evaluated and validated using the NinaPro DB6 dataset, which contains a rich variety of sEMG data related to hand and finger movements. Our findings demonstrate that the proposed approach outperforms a prior art method, showcasing its effectiveness in accurately identifying the onset and offset points of muscle activation during gestures. These results are of utmost importance for applications such as human-robot interaction, rehabilitation, and assistive technology where precise gesture recognition is a critical component.