Abstract:

In recent years, there has been rapid development in the research area of the car's driving assistance. Deep learning was used to solve different problems, such as object detection, visual recognition, speech recognition and handwriting recognition and was achieved a very good performance. In deep learning, Convolutional Neural Networks (ConvNets or CNNs) are found to give the most accurate results, in solving object detection problems. In this paper, we'll go into summarizing some of the most important deep learning models used for object detection tasks over this last recent year, then we present an approach where we detect roadsides, then we seek objects located on the road area to prevent driver. As the state of the driver is very important information, we try to detect driver's drowsiness. We use a camera with an algorithm to calculate the eye aspect ratio. Finally, we evaluated the three modules of proposed system using our collected data-set.

Publisher's Version