Publication Date: 9/9/2026

Event: 6th DriveX Workshop in conjunction with ECCV 2026

Reference: pp. 1-16, 2026

Authors: Sparsh Garg, NEC Laboratories America, Inc.; Yi-Wen Chen, NEC Laboratories America, Inc.; Vijay Kumar B G, NEC Laboratories America, Inc.; Abhishek Aich, NEC Laboratories America, Inc.

Abstract: While modern autonomous driving systems excel at perception tasks such as object detection and trajectory prediction, they lack the high-level causal reasoning required to interpret traffic accidents.In particular, determining responsibility, such as identifying who is at fault and which traffic rule was violated, remains largely unexplored in current benchmarks. To this end, we introduce CAViAR (Causal Accident Video and Incident Analysis Repository), a human-annotateddashcam benchmark comprising 2,249 real-world accident videos collected from CarCrashDataset (CCD) and Nexar. Each video is annotated with structured labels spanning environmental conditions, accident type, causal explanation, apparent At-Fault Agent, affected agent, and apparent rule-violation category. We benchmark state-of-the-art vision-language models (VLMs), including Cosmos-Reason2, Qwen3-VL, andInternVL3. Once class imbalance is accounted for with majority/randombaselines and balanced metrics, perceptual competence is uneven–lighting

Publication Link: https://openreview.net/forum?id=5qPVXRwUBJ , https://arxiv.org/pdf/2608.19380