Padik - Pedestrian-Driver Interaction Dataset in Heterogeneous, Unstructured Traffic Environments

India ranks first in road accidents and fatalities globally. To address this problem, It is crucial to understand the nature of interaction between the pedestrian and the driver. This paper presents Padik, a comprehensive dataset gathered to explore the behavioural patterns of pedestrians and drivers during crossing events in unstructured, heterogeneous, and dynamic traffic environments. Multiple factors such as weather conditions, time of day, pedestrian gender, group dynamics, and designated crossing areas are annotated. The Padik dataset aims to enhance the accuracy of Advanced Driver Assistance Systems (ADAS) and Autonomous Vehicles (AV) by providing detailed data and information on pedestrian-driver interactions in heterogenous traffic environments. The outputs of this research are essential for improving road safety technologies, creating sustainable urban planning, and eventually reducing the number of injuries and fatalities on roads.