A hierarchical CNN-Tracker Approach for Scene Classification for Autonomous Vehicles

Driver safety in automobiles has been a critical concern since the inception of on-road vehicles. Over the years, leading researchers and Original Equipment Manufacturers (OEMs) have diligently worked to address safety issues by developing various advanced safety systems. According to the World Health Organization, driver negligence results in approximately 1.19 million fatalities annually. Furthermore, these incidents incur a global economic cost of around $518 billion every year, accounting for about 1–2% of the total gross domestic product of all countries. In response to these high casualty rates and increasing consumer demand, OEMs have intensified their efforts to develop Advanced Driver Assistance Systems (ADAS). These systems aim to enhance occupant protection and reduce fatalities, establishing new standards in automotive safety.