Phát hiện ngủ gật khi lái xe ô tô dựa trên kỹ thuật học sâu

Abstract

Drowsiness while driving is one of the most common causes of traffic accidents around the world, the results of which damage, injury and death can be permanent. In this paper, we propose a doze detection method using deep learning network developed based on InceptionV3 and DenseNet networks. We take advantage of deep learning network and forward learning approach to train recommender networks on our dataset. This not only solves the problem of data set limitations, but also gives the comparison results in terms of time as well as the accuracy of neural networks. Experimental results of the proposed method can achieve accuracy up to 98%.