Communication Through Real-Time Video Oculography Using Face Landmark Detection
R Rakshita · 2018
An alternate method of communication for Amyotropic Lateral Sclerosis (ALS) patients using realtime video oculography is introduced in this paper. Proposed method is user-friendly and assist the patients to communicate their basic requirements. The algorithm can be used to address large number of outputs through various input eye wink combinations. The system is built using openCV and dlib. Eye wink is detected by computing Eye Aspect Ratio(EAR) using face landmarks obtained from dlib. Pattern in eye winks is considered to provide corresponding audio and visual outputs. This method can easily differentiate between voluntary and involuntary blinks with better accuracy.