DIRS21-Dataset for Indian Road Scenarios

We propose a novel high-resolution dataset named, “Dataset for Indian Road Scenarios (DIRS21)” for developing perception systems for advanced driver assistance systems.

For data acquisition, we utilized 12- megapixel camera with an f/1.8 aperture with 1920 X 1080 pixels resolution. This dataset consists of 5093 images, those were collected under various scenes, weather, season, and illumination conditions. Each image is labelled with the name of the class to which it belongs. Our work considers the following seven classes of objects: Pedestrian, Rider, Car, Bus, Truck, NMV (Non-Motorized Vehicle), and EA (Electric Auto). Moreover, In our dataset, we consider (i) sedan and SUV in the same class called ‘Car’, (ii) minibus and Volvo bus are together grouped in the class ’Bus’, (iii) Cyclists and motorcyclists are placed in the class ’Rider’.