Fatigue driving detection with modified ada-boost and fuzzy algorithm

Abstract

Facial features is the most important and obvious characteristics for the fatigue driving detection. This paper uses the modified Ada-Boost algorithm to detect face and to locate eyes and mouth precisely. The adaptive threshold is used to extract the characteristics of the eyes and mouth status. At last, fuzzy algorithm is used to judge the fatigue status which combined with PERCLOS rules. Experiments show that the proposed method has stronger robustness, faster speed, more accurate precision and meet the real-time demand.