SafeAug: Safety-Critical Driving Data Augmentation from Naturalistic Datasets

TL;DR

This paper introduces SafeAug, a data augmentation framework that enhances naturalistic driving datasets with safety-critical scenarios using vehicle detection, depth estimation, and 3D transformations, improving self-driving model training.

Contribution

SafeAug is a novel framework that generates authentic safety-critical driving data from naturalistic datasets, bridging the gap between real and simulated data for safer self-driving algorithms.

Findings

Augmented data improves self-driving model performance.

Method outperforms SMOGN and importance sampling baselines.

Maintains high image authenticity in safety-critical scenarios.

Abstract

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.