SafeAug: Safety-Critical Driving Data Augmentation from Naturalistic Datasets
This addresses the data authenticity gap for developing safer self-driving systems, though it is an incremental improvement over existing augmentation methods.
The paper tackles the scarcity of authentic safety-critical driving data for self-driving algorithms by proposing a framework to augment such data from naturalistic datasets, resulting in a downstream algorithm that performs superiorly compared to baselines like SMOGN and importance sampling.
Safety-critical driving data is crucial for developing safe and trustworthy self-driving algorithms. Due to the scarcity of safety-critical data in naturalistic datasets, current approaches primarily utilize simulated or artificially generated images. However, there remains a gap in authenticity between these generated images and naturalistic ones. We propose a novel framework to augment the safety-critical driving data from the naturalistic dataset to address this issue. In this framework, we first detect vehicles using YOLOv5, followed by depth estimation and 3D transformation to simulate vehicle proximity and critical driving scenarios better. This allows for targeted modification of vehicle dynamics data to reflect potentially hazardous situations. Compared to the simulated or artificially generated data, our augmentation methods can generate safety-critical driving data with minimal compromise on image authenticity. Experiments using KITTI datasets demonstrate that a downstream self-driving algorithm trained on this augmented dataset performs superiorly compared to the baselines, which include SMOGN and importance sampling.