Real-Time Eye-Tracking-Based Cursor Control for Intelligent Human-Computer Interaction

This paper suggests an affordable, webcam- based eye-tracking and blink-detection system to enable hands- free human-computer interaction. Building on MediaPipe for facial landmark detection, OpenCV for image processing, and PyAutoGUI for control, the system converts gaze direction into cursor movement and identifies purposeful blinks for click operations. The system is shown to have very high blink detection accuracy and real-time response, backed by a calibrated Eye Aspect Ratio (EAR) value of 0.28. Blinks with a duration greater than 300 ms are assumed to be intentional, allowing precise click simulation while keeping false positives low. This solution introduces increased accessibility for motor disabled users without the need for special hardware. The system accommodates natural, real-time cursor control and is appropriate for use in smart environments, virtual reality, and assistive technology.