Published Paper Details:

Every year many people lose their lives due to road accidents. The cause for road accidents may be many reasons and one of the common reasons is the drowsiness of a driver. Micro drowsiness or fatigue can be a major cause to a huge accident. Thus, initial detection of fatigue before a critical situation may occur, is the ongoing research of this paper. A real-time eye-blink and yawn detection algorithm is proposed in this paper. The recent landmark detectors are trained on various datasets, with varying light and facial expressions and tested against robustness and intelligence with respect to a camera. The first step is to detect the facial region and capture video sequences using computer vision techniques. In the next step, the facial landmark detectors are used for locating the eyes and mouth region of the driver which are studied using the eye aspect ratio(EAR), which is a single scalar quantity to detect the eye and mouth opening in each frame. We show that the facial landmarks are detected precisely to show the level of eye and mouth opening. Finally, in collaboration, the driver state that is, asleep or awake, is determined and a warning alarm is played if sleep is detected. These various experiments prove a high accuracy against the proposed idea.