Deep Learning Approach Based on Residual Neural Network and SVM Classifier for Driver’s Distraction Detection

In the last decade, distraction detection of a driver gained lot significance due to increases in number accidents. Many solutions, such as feature based, statistical, holistic, etc., have been proposed solve this problem. With advent high processing power at cheaper costs, deep learning-based techniques shown promising results. The study proposes ReSVM, an approach combining features ResNet-50 with SVM classifier, for driver. ReSVM is compared six state-of-the-art approaches on four datasets, namely: State Farm Distracted Driver Detection, Boston University, DrivFace, and FT-UMT. Experiments demonstrate that outperforms existing achieves classification accuracy 95.5%. also compares its variants aforementioned datasets.