BlinkSpeller: Dynamic EAR-Based Eye Blink Morse Code Communication System with Predictive Text and Speech Synthesis

Abstract: Assistive communication technologies are crucial to the communication of people who cannot use traditional input devices because of their severe motor impairments. In this paper, a low cost, web-based communication system is presented that uses eye blink detection to provide text and speech output in real time. The system uses MediaPipe Face Mesh for facial landmark detection and calculates the Eye Aspect Ratio (EAR) to detect blinks. A custom algorithm categorizes blinks into short and long blinks (Morse code) and adds noise filtering, head motion compensation and gaze stabilization to make it more robust. A special decoding engine is used to decode the generated Morse sequences into alphanumeric characters and control commands. The system is built on browser-based APIs such as the Webcam API, SpeechSynthesis API, and localStorage, allowing for real-time data processing, speech generation, and storage. A word prediction module increases the efficiency of input. The proposed system shows high reliability, low latency and high accessibility, which can provide an effective solution for hands-free human–computer interaction.

Keywords: Assistive Communication System, Eye Blink Detection, MediaPipe Face Mesh, Eye Aspect Ratio (EAR), Morse Code Communication, Speech Synthesis, Computer Vision, Human–Computer Interaction