[This article belongs to Volume - 54, Issue - 06]
Gongcheng Kexue Yu Jishu/Advanced Engineering Science
Journal ID : AES-10-08-2022-268

Title : Automated driver drowsy system Using Eye Landmark and Gaze Estimation
Bharathi R, Sumathi MS, Bhagya R, ShivaPrasad Yadav SG, Chaitra Kumari,

Abstract : Vehicle crashes on road increases day by day. According to the ASIRT on average minimum 3287 people lose their life in road accident. There are several parameters which lead to the road accident in which driver activity is one of the significant parameters. Due to sleep disorder and driver dullness during driving will lead to the road accidents. To overcome this problem and to monitor the driver activity a driver drowsy alert system is designed. A number of researches are conducted on the alert system to give best system response. In proposed work we implemented a driver drowsy monitoring system, in which a high-resolution camera is fixed in front of the driver which continuously record and monitor the driver eye movement. The module is designed by using Histogram of Oriented gradients (HOG) and Haar cascade face detection algorithm. Regression tree method is included to point out the eye region and its position. Once the eye region is obtained pupil is detected and tracked using gradient method and kalman filter. The module is implemented in visual studio 2015, by using C++, Open CV and Dlib library function. The code is ported on Renesas R CAR E2 SILK platform. The system generates the warning signal when driver close his eye more than predefined and when he/she distracted from the proper position.