Tracking and Classification of Head Movement for Augmentative and Alternative Communication Systems

Area of Science:

• Computer Science
• Rehabilitation Engineering
• Human-Computer Interaction

Background:

• Assistive technologies aid computer use for motor disabilities.
• Individuals with severe involuntary movements have limited options.
• Conventional input methods are challenging for users with significant motor impairments.

Purpose of the Study:

• To develop and evaluate a head-motion-controlled virtual keyboard for users with motor disabilities.
• To assess the performance of a personalized Hidden Markov Model (HMM) classifier.
• To investigate the system's usability for individuals with cerebral palsy.

Main Methods:

• A conventional webcam captures and tracks user head motion.
• Functional movements are modeled and recognized for virtual keyboard input.
• A flexible, personalized Hidden Markov Model (HMM) classifier is proposed.
• Performance is compared against a position threshold classifier.

Main Results:

• HMM-based classifiers performed comparably or better than position threshold methods in unimpaired users.
• Motion segmentation and interpretation modules showed sensitivity to involuntary movements.
• Participants with cerebral palsy experienced challenges with system sensitivity to their involuntary movements.

Conclusions:

• The proposed head-tracking virtual keyboard offers a personalized assistive solution.
• Further refinement is needed to robustly handle involuntary movements in users with cerebral palsy.
• The system demonstrates potential for improving computer accessibility for individuals with motor impairments.