Fig. 3

Flow diagram of the experimental protocol. The experiment was divided into two phases: data collection (offline phase) and online assessment. During the offline phase, HD-EMG data was collected and features were extracted. After training the classifiers for different number of channels selected using SFS, CIRC and ALL, the offline classification accuracy was calculated using testing data. Once the offline analysis was completed, three electrode configurations (SFS, CIRC and ALL) were evaluated during an online gesture recognition task