Eye Gesture Analysis with Head Movement for Advanced Driver Assistance Systems

Road traffic accidents are a major cause of death worldwide. In an attempt to reduce accidents, some research efforts have focused on creating Advanced Driver Assistance Systems (ADAS) able to detect vehicle, driver and environmental conditions and to use this information to identify cues for potential accidents. This paper presents continued work on a novel Non-intrusive Intelligent Driver Assistance and Safety System (Ni-DASS) for assessing driver point of regard within vehicles. It uses an onboard CCD camera to observe the driver’s face. A template matching approach is used to compare the driver’s eye-gaze pattern with a set of eye-gesture templates of the driver looking at different focal points within the vehicle. The windscreen is divided into cells and comparison of the driver’s eye-gaze pattern with templates of a driver’s eyes looking at each cell is used to determine the driver’s point of regard on the windscreen. Results indicate that the proposed technique could be useful in situations where low resolution estimates of driver point of regard are adequate. For instance, To allow ADAS systems to alert the driver if he/she has positively failed to observe a hazard.