Efficient strategies for finger movement classification using surface electromyogram signals.

Area of Science:

• Biomedical Engineering
• Pattern Recognition
• Signal Processing

Background:

• Surface electromyogram (sEMG) signals are crucial for hand and finger gesture recognition.
• Accurate classification of finger movements is a significant challenge in human-computer interaction and prosthetics.

Purpose of the Study:

• To propose and evaluate four distinct techniques for classifying finger movements based on sEMG signals.
• To identify the most effective method for high-accuracy sEMG-based finger movement classification.

Main Methods:

• Technique 1: Dynamic graph construction and graph entropy.
• Technique 2: Dimensionality reduction (LTSA, LLC) with Evolutionary Algorithms (EA), Bayesian Belief Networks (BBN), and Extreme Learning Machines (ELM) (EA-BBN-ELM).
• Technique 3: Differential Entropy (DE), Higher-Order Fuzzy Cognitive Maps (HFCM), Empirical Wavelet Transformation (EWT) (DE-FCM-EWT).
• Technique 4: Local Mean Decomposition (LMD), Fuzzy C-Means clustering, and Least Squares Support Vector Machine (LS-SVM).

Main Results:

• The LMD-fuzzy C-means clustering with LS-SVM achieved the highest classification accuracy of 98.5%.
• The DE-FCM-EWT hybrid model with SVM classifier yielded the second-best accuracy at 98.21%.
• The LTSA-based EA-BBN-ELM model achieved a classification accuracy of 97.57%.

Conclusions:

• The LMD-fuzzy C-means clustering combined with LS-SVM demonstrates superior performance for sEMG-based finger movement classification.
• Hybrid models integrating advanced signal processing and machine learning techniques show significant promise in this field.
• The study provides valuable insights into optimizing sEMG signal analysis for enhanced gesture recognition systems.