Enregistré dans:
| Auteurs principaux: | , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2025
|
| Sujets: | |
| Accès en ligne: | https://arxiv.org/abs/2510.22260 |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866909870039498752 |
|---|---|
| author | Zhao, Tianhao Zou, Yiyang Mao, Zihao Xiao, Peilun Huang, Yulin Yang, Hongda Li, Yuxuan Li, Qun Wu, Guobin Lin, Yutian |
| author_facet | Zhao, Tianhao Zou, Yiyang Mao, Zihao Xiao, Peilun Huang, Yulin Yang, Hongda Li, Yuxuan Li, Qun Wu, Guobin Lin, Yutian |
| contents | Accident anticipation aims to predict potential collisions in an online manner, enabling timely alerts to enhance road safety. Existing methods typically predict frame-level risk scores as indicators of hazard. However, these approaches rely on ambiguous binary supervision (labeling all frames in accident videos as positive) despite the fact that risk varies continuously over time, leading to unreliable learning and false alarms. To address this, we propose a novel paradigm that shifts the prediction target from current-frame risk scoring to directly estimating accident scores at multiple future time steps (e.g., 0.1s-2.0s ahead), leveraging precisely annotated accident timestamps as supervision. Our method employs a snippet-level encoder to jointly model spatial and temporal dynamics, and a Transformer-based temporal decoder that predicts accident scores for all future horizons simultaneously using dedicated temporal queries. Furthermore, we introduce a refined evaluation protocol that reports Time-to-Accident (TTA) and recall (evaluated at multiple pre-accident intervals (0.5s, 1.0s, and 1.5s)) only when the false alarm rate (FAR) remains within an acceptable range, ensuring practical relevance. Experiments show that our method achieves superior performance in both recall and TTA under realistic FAR constraints. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_22260 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Accident Anticipation via Temporal Occurrence Prediction Zhao, Tianhao Zou, Yiyang Mao, Zihao Xiao, Peilun Huang, Yulin Yang, Hongda Li, Yuxuan Li, Qun Wu, Guobin Lin, Yutian Computer Vision and Pattern Recognition Accident anticipation aims to predict potential collisions in an online manner, enabling timely alerts to enhance road safety. Existing methods typically predict frame-level risk scores as indicators of hazard. However, these approaches rely on ambiguous binary supervision (labeling all frames in accident videos as positive) despite the fact that risk varies continuously over time, leading to unreliable learning and false alarms. To address this, we propose a novel paradigm that shifts the prediction target from current-frame risk scoring to directly estimating accident scores at multiple future time steps (e.g., 0.1s-2.0s ahead), leveraging precisely annotated accident timestamps as supervision. Our method employs a snippet-level encoder to jointly model spatial and temporal dynamics, and a Transformer-based temporal decoder that predicts accident scores for all future horizons simultaneously using dedicated temporal queries. Furthermore, we introduce a refined evaluation protocol that reports Time-to-Accident (TTA) and recall (evaluated at multiple pre-accident intervals (0.5s, 1.0s, and 1.5s)) only when the false alarm rate (FAR) remains within an acceptable range, ensuring practical relevance. Experiments show that our method achieves superior performance in both recall and TTA under realistic FAR constraints. |
| title | Accident Anticipation via Temporal Occurrence Prediction |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2510.22260 |