Driver blink measurement by the motion picture processing and its application to drowsiness detection

There is a requirement for driver assistance systems in ITS to be made driver-friendly. Therefore, unusual driver states (drowsiness, inattentive driving, excessive concentration, etc.) must be detected. A method for measuring the blinking of a driver in real time by motion picture processing was developed. Moreover, a method for presuming consciousness degradation from the change in blink duration was developed. The waveform of blinking differs for individual people and differs over time for the same person. The method developed for processing blink waveforms is robust against the influence of differences among individuals. The blink extraction rate obtained with ten subjects is considerably improved from 76% to 95%, and we have the prospect of presuming degree of consciousness degradation from these results.