Deteksi Kantuk pada Pengemudi Berdasarkan Penginderaan Wajah Menggunakan PCA dan SVM

Abstract

—Drowsiness while driving is one of the main causes of traffic accidents it affects the level of focus of the driver. Therefore, we need an automatic drowsiness detection mechanism for the driver to provide a warning or alarm so that an accident can be avoided. In this study, we design and simulate a system to detect drowsiness through the driver’s yawn expression. The acquisition is made by recording the face from two shooting points including the dashboard and front mirrors in the car. From the video recording, then it is taken into several images with a size of 128x82 pixels which are used as training and testing data. This image is then processed using Principal Component Analysis (PCA) for feature extraction and classified using a Support Vector Machine (SVM). From the tests carried out, the system generates the highest accuracy of 98%. This best performance is obtained by SVM with polynomial kernel in the camera position on the dashboard. Meanwhile, based on compression testing, the image that can still meet system requirements is 25% of the original size. It is hoped that the proposed drowsiness detection method in this study can be applied for real-time drowsiness detection in vehicles.