Rectus Femoris and Gastrocnemius EMG Driven Cheonjiin Speller for Korean Text Input

Area of Science:

• Biomedical Engineering
• Rehabilitation Technology
• Human-Computer Interaction

Background:

• Individuals with restricted hand mobility face challenges in text communication.
• Existing assistive technologies may have limitations in terms of accessibility and computational load.

Purpose of the Study:

• To develop and evaluate a surface electromyography (sEMG)-based Cheonjiin speller system.
• To enable efficient and accessible text input for users with motor impairments.
• To assess the system's performance and feasibility for real-time application.

Main Methods:

• Recorded two-channel sEMG signals from leg muscles.
• Processed signals in real-time using filtering and rectification.
• Extracted time-domain features (RMS, SSC, peak amplitude) for classification.
• Implemented a directional control framework with a Korean keyboard layout.

Main Results:

• Achieved 90.0% accuracy for discrete command recognition.
• Reached 88.65% accuracy for continuous Korean word and sentence input.
• Attained an average information transfer rate (ITR) of 96.19 bits/min.
• Demonstrated high recognition performance using low-computation features without deep learning.

Conclusions:

• The sEMG-based Cheonjiin speller system is highly operable, accessible, and practical for users with motor impairments.
• The system's lightweight architecture and efficient processing make it suitable for resource-limited environments.
• It holds significant potential for integration into augmentative and alternative communication (AAC) systems and wearable assistive technologies.