A Novel Magnetometer Array-based wearable system for ASL gesture recognition.

Area of Science:

• Robotics and Human-Computer Interaction
• Biomedical Engineering
• Signal Processing

Background:

• Hand gesture classification is crucial for sign language recognition (SLR) systems aiding individuals with hearing and speech impairments.
• Existing methods like surface-based electromyography (sEMG), inertial measurement units (IMUs), flex sensors, and video-based systems have limitations including computational intensity, bulkiness, or restricted user movement.
• The need for a highly discriminative, non-intrusive, and cost-effective solution for American Sign Language (ASL) recognition is significant.

Purpose of the Study:

• To propose a novel, precision-preserved static hand gesture classification system for ASL recognition.
• To develop a cost-effective and minimally intrusive wearable system for SLR.
• To evaluate the accuracy and robustness of the proposed system.

Main Methods:

• Development of a wearable system utilizing an array of magnetometers for static hand gesture classification.
• Implementation of the K-Nearest Neighbors (KNN) classification model to process magnetometer data.
• Design considerations focused on minimizing electronic coverage on the hand and optimizing system cost.

Main Results:

• Achieved an average accuracy of 98.60% for classifying ASL alphabets.
• Attained an average accuracy of 94.07% for classifying ASL digits.
• Demonstrated robust classification results with a system that occupies negligible space on the user's hand.

Conclusions:

• The proposed magnetometer array-based wearable system offers a cost-effective, reliable, and robust solution for ASL recognition.
• The system's non-intrusive design and high accuracy make it suitable for practical adoption.
• This approach addresses the limitations of existing SLR technologies, paving the way for improved communication tools for the hearing-deprived community.