Eye-Gaze Controlled Virtual Keyboard Using Blink Detection for Accessible Hands-Free Communication

The inability to interact with traditional input devices like keyboards poses significant challenges for individuals with motor impairments. Existing assistive technologies often rely on cumbersome and expensive hardware solutions, which are not accessible to everyone. This paper addresses the problem by presenting an affordable and accessible gaze-controlled virtual keyboard that uses only a standard webcam. The project utilizes computer vision algorithms, specifically dlib's face detection and shape prediction, to detect facial landmarks and calculate the Eye Aspect Ratio (EAR), which enables accurate eye-blink detection. The system is designed to dynamically highlight letter groups on a virtual keyboard, allowing users to select keys through a series of eye blinks. The keyboard switches between letter groups and individual letters, facilitating text entry without requiring physical input. The proposed solution demonstrates how computer vision techniques can be used to enhance accessibility and independence for users with severe motor disabilities by providing a practical and user-friendly interface.