Algorithm 1: Pseudocode of the introduced Frustum154.

Input: sEMG dataset

Output: Results.

00: Read each sEMG signal from datasets.

01: Merge the channels

02: Apply for TQWT decomposition to calculate 153 subbands.

03: Extract statistical and texture features from sEMG signal and subbands.

04: Obtain a 154-feature vector with a length of 414 (=384 + 30).

05: Calculate the misclassification rate of each feature vector.

06: Choose the best 20 feature vectors using the calculated misclassification rates.

07: Merge the 20 feature vectors selected to create final feature vector.

08: Choose the most informative feature by applying INCA.

09: Classify the chosen features using kNN or SVM with 10-fold cross-validation.