SEMG Feature Extraction Based on StockwellTransform Improves Hand MovementRecognition Accuracy

Area of Science:

• Biomedical Engineering
• Signal Processing
• Machine Learning

Background:

• Surface electromyography (SEMG) is crucial for understanding muscle activity.
• Traditional time or frequency domain analysis of SEMG signals has limitations in information extraction.
• Time-frequency analysis offers a more comprehensive approach to SEMG signal interpretation.

Purpose of the Study:

• To propose a novel time-frequency analysis method using the Stockwell transform (S-transform) for improved hand movement recognition from SEMG signals.
• To enhance the accuracy and efficiency of SEMG-based hand movement classification.

Main Methods:

• Feature extraction from forearm SEMG signals using the S-transform.
• Dimensionality reduction of extracted features via Principal Component Analysis (PCA) for computational efficiency.
• Hand movement recognition using an Artificial Neural Network (ANN) Multilayer Perceptron (MLP) classifier.

Main Results:

• The S-transform-based feature extraction significantly improved class separability.
• The proposed method demonstrated higher hand movement recognition accuracy compared to wavelet transform and power spectral density methods.
• PCA effectively reduced feature vector dimensionality, improving classifier speed.

Conclusions:

• The S-transform is a powerful tool for extracting informative features from SEMG signals.
• This approach enhances the accuracy of hand movement recognition for SEMG-based applications.
• The combination of S-transform, PCA, and ANN offers a robust solution for SEMG pattern recognition.