An Adaptive and Robust Implementation of a Fatigue Detection System

Advanced driver assistance systems are the need of the hour. One component, the fatigue or drowsy driver detection system has been an ongoing development effort. This paper aims to propose one such cost effective yet accurate system which localizes the areas of the eyes and the mouth to determine whether a person is experiencing fatigue by computing certain fatigue metrics. The paper also aims to show how the method of binarization along with a few image processing algorithms can be a strong ensemble for such a system.