Driver Distraction Identification using Multiple Machine Learning Approaches

Abstract

According to the preceding year’s road statistical report emphasize that the prime reasons of mortal road accidents are because of drowsy or distracted state of driver. Recognition of such critical states of driver at its initial phase with higher accuracy can rescue several precious lives. To satisfy this demand, we have analyzed the five different classifier’s i.e. Fuzzy min-max, Decision tree, K- Nearest Neighbor’s, Linear Support Vector Machine and VGG-16 neural network. Among these classifier’s, VGG-16 has given outstanding result with accuracy of 96.4 % on validation data but lagged in the terms training time.