CycleCrash: A Dataset of Bicycle Collision Videos for Collision Prediction and Analysis
This work addresses the underrepresentation of cyclist collisions in self-driving research by providing a dedicated dataset and method, which is an incremental step towards improving safety for cyclists.
This paper introduces CycleCrash, a new dataset of 3,000 dashcam videos (436,347 frames) focusing on bicycle collisions and interactions to improve cyclist safety in self-driving research. The dataset supports 9 different collision prediction and classification tasks and is accompanied by VidNeXt, a novel method for temporal video analysis.
Self-driving research often underrepresents cyclist collisions and safety. To address this, we present CycleCrash, a novel dataset consisting of 3,000 dashcam videos with 436,347 frames that capture cyclists in a range of critical situations, from collisions to safe interactions. This dataset enables 9 different cyclist collision prediction and classification tasks focusing on potentially hazardous conditions for cyclists and is annotated with collision-related, cyclist-related, and scene-related labels. Next, we propose VidNeXt, a novel method that leverages a ConvNeXt spatial encoder and a non-stationary transformer to capture the temporal dynamics of videos for the tasks defined in our dataset. To demonstrate the effectiveness of our method and create additional baselines on CycleCrash, we apply and compare 7 models along with a detailed ablation. We release the dataset and code at https://github.com/DeSinister/CycleCrash/ .