Authors:
Jon Ander Iñiguez de Gordoa
1
;
2
;
Martín Hormaetxea
1
;
Marcos Nieto
1
;
Gorka Vélez
1
and
Andoni Mujika
2
Affiliations:
1
Vicomtech Foundation, Basque Research and Technology Alliance (BRTA), Mikeletegi 57, 20009, Donostia-San Sebastián, Spain
;
2
University of the Basque Country (UPV/EHU), Donostia-San Sebastián, Spain
Keyword(s):
Synthetic Data, Simulation, Unreal Engine, Diversity, Mobility Aids, Fisheye, ADAS.
Abstract:
This work presents DiverSim, a highly customizable simulation tool designed for the generation of diverse synthetic datasets of vulnerable road users to address key challenges in pedestrian detection for Advanced Driver Assistance Systems (ADAS). Although recent Deep Learning models have advanced pedestrian detection, their performance still depends on the diversity and inclusivity of training data. DiverSim, developed on Unreal Engine 5, allows users to control various environmental conditions and pedestrian characteristics, including age, gender, ethnicity and mobility aids. The tool features a highly customizable virtual fisheye camera and a Python API for easy configuration and automated data annotation in the ASAM OpenLABEL format. Our experiments demonstrate DiverSim’s capability to evaluate pedestrian detection models across diverse user profiles, revealing potential biases in current state-of-the-art models. By making both the simulator and Python API open source, DiverSim ai
ms to contribute to fairer and more effective AI solutions in the field of transportation safety.
(More)