Hand Gesture Recognition and Biometric Authentication Using a Multi-day Dataset

Hand-gesture recognition (HGR) is one of the major applications of electromyography (EMG), specifically for controlling functional prosthetic hands. Recently, another application i.e. the EMG-based biometrics has found growing research interest due to its potential of addressing some conventional biometric limitations. It has been observed that for EMG-based applications, the translation of laboratory research to real-life applications suffers from two major limitations: 1) a small subject pool, and 2) limited to single-session data recordings. In this study, forearm, and wrist EMG data were collected from 43 participants over three different days with long separation (Days 1, 8, and 29) while they performed static hand/wrist gestures. The HGR evaluation resulted in a mean AUC of 0.948 +/- 0.018 and 0.941 +/- 0.021 for forearm data and wrist data, respectively. The biometric evaluation resulted in a mean EER of 0.028 +/- 0.007 and 0.038 +/- 0.006 for forearm data and wrist data, respectively. These results were comparable to the widely used Ninapro database DB2. The large-sample multi-day dataset would facilitate further research on EMG-based HGR and biometric applications.