Progressive Risk Estimation for Accident Anticipation

Samet Hicsonmez; Eray Çakar; Nermin Samet; Fatma Güney

cs.CVSep 26, 2026

Progressive Risk Estimation for Accident Anticipation

Authors: Samet Hicsonmez, Eray Çakar, Nermin Samet, Fatma Güney

Organizations: Independent Researcher, Paris, France · Department of Computer Engineering, KUIS AI, Koç University, Istanbul, Turkey · Valeo.ai, Paris, France

Abstract

Accident anticipation aims to recognize anomalous driving cues before a crash while avoiding false alarms during normal driving. Existing approaches typically formulate this task as binary classification, focusing on whether an accident will occur rather than when it will occur. We propose PRE-ACT, a framework that models accident risk as a continuously evolving signal that increases as the crash approaches. By explicitly enforcing temporal ordering and distance-to-accident awareness, our method progressively raises risk while suppressing premature alarms, leading to significant improvements on MM-AU subsets and Nexar. We further introduce a Separation Score to evaluate the global behavior of predicted risk curves beyond local temporal windows. Code and visualizations are available at https://github.com/giddyyupp/PRE-ACT.

Figures & tables

Panel of Figure 1 — Figure 1: PRE-ACT : From binary accident prediction to progressive risk anticipation. Unlike the dominant paradigm of formulating accident anticipation as a binary classification of whether an accident will occur (baseline), we model risk as a temporally evolving signal that reflects when an accident is likely to occur. By explicitly enforcing temporal order and distance-to-crash awareness, our model reduces false alarms while progressively increasing risk as abnormal cues emerge. Scores are overlaid on frames as an alpha channel, with blue indicating safe and red indicating risky predictions.

Panel of Figure 2 — Figure 2: Overview . We define the period before the fixed anticipation horizon $t_{h}$ as the safe zone, and the interval from $t_{h}$ to the accident time $t_{\text{acc}}$ as the critical zone with progressively increasing risk. During training, we extract spatio-temporal features using a video foundation model. Based on these features, the model classifies clips as safe or critical, assigns a continuous risk score, and ranks them using preference head outputs, encouraging higher scores for critical clips closer to the accident. Dashed arrows indicate components used only during training.

Panel of Figure 3 — Figure 3: Varying $\alpha$ . Continuous risk curves with fast-increasing (red side) and slow-increasing (blue side) behavior obtained by varying $\alpha$ . The linear case is shown with a dashed line.

Panel of Figure 4 — Figure 4: Problem with current evaluation metrics. Standard AUC evaluates predictions only within local temporal windows: negatives are sampled around anomaly onset ( $t_{ai}-0.5$ ) and positives close to the accident ( $t_{acc}-2$ ), illustrated by the yellow hatched regions. As a result, the cyan , magenta , and green curves achieve similarly high AUC (near 1) and identical TTA, since they rise relative to the pre-anomaly region at comparable times ( orange line). However, their global behavior differs significantly. The cyan curve produces early false alarms before $t_{ai}$ , while the magenta curve shows weak progression toward the accident. The green curve exhibits the desired behavior: low risk before $t_{ai}$ , a clear increase after anomaly onset, and steadily rising risk as $t_{acc}$ ( red line) approaches.

	Method	AUC ${}_{0.0s}^{0.1}$	AUC ${}_{0.5s}^{0.1}$	AUC ${}_{1.0s}^{0.1}$	AUC ${}_{1.5s}^{0.1}$	mAUC 0.1	mTTA 0.1
CAP	CAP [ 26 ]	0.042	0.040	0.030	0.037	0.036	0.637
	DRIVE [ 43 ]	0.129	0.117	0.108	0.123	0.116	0.395
	DSTA [ 9 ]	0.559	0.386	0.282	0.191	0.286	0.804
	GSC [ 10 ]	0.609	0.418	0.297	0.199	0.305	0.817
	TOP [ 11 ]	0.838	0.675	0.398	0.214	0.429	0.864
	PRE-ACT	0.887	0.775	0.465	0.201	0.481	1.125

Table 1: Results on MM-AU. We report results on the CAP [ 26 ] (top) and DADA [ 39 ] (bottom) splits of MM-AU [ 12 ] . Our method consistently outperforms prior work across almost all metrics by substantial margins, with improvements becoming more pronounced closer to the accident.

Method	mAP
Runner-up	0.875
Winner	0.885
PRE-ACT	0.895

Table 2: Results on Nexar.

Train	Test	AUC ${}_{0.0s}^{0.1}$	AUC ${}_{0.5s}^{0.1}$	AUC ${}_{1.0s}^{0.1}$	AUC ${}_{1.5s}^{0.1}$	mAUC 0.1	mTTA 0.1	$s$
CAP	DADA	0.897	0.799	0.413	0.178	0.464	1.155	0.760
diff.		0.036	0.048	0.015	0.008	0.024	0.086	0.020
	Nexar	-	0.529	0.336	0.218	0.361	1.020	0.650
diff.		-	0.082	0.236	0.151	0.149	0.265	0.049

Table 3: Cross-dataset generalization. We train our model on the largest subset, CAP, and evaluate it on DADA and Nexar. As indicated by the performance differences (diff.; increase and decrease ) relative to models trained on each dataset’s own training set, our model generalizes well to DADA, with a slight performance drop on Nexar due to a larger domain gap.

BCE	Prog.	Pref.	AUC ${}_{0.0s}^{0.1}$	AUC ${}_{0.5s}^{0.1}$	AUC ${}_{1.0s}^{0.1}$	AUC ${}_{1.5s}^{0.1}$	mAUC 0.1	mTTA 0.1	$s$
✓	✗	✗	0.807	0.690	0.392	0.176	0.419	1.105	0.694
✓	✓	✗	0.884	0.773	0.463	0.197	0.477	1.128	0.723
✓	✓	✓	0.887	0.775	0.465	0.201	0.481	1.125	0.724

Table 4: Ablation on progress and preference estimation. Compared to the BCE baseline, progress estimation (Prog.) significantly improves performance across all metrics. Adding preference estimation (Pref.) provides slight additional gains, achieving the best overall results.

Increase	$\alpha$	AUC ${}_{0.0s}^{0.1}$	AUC ${}_{0.5s}^{0.1}$	AUC ${}_{1.0s}^{0.1}$	AUC ${}_{1.5s}^{0.1}$	mAUC 0.1	mTTA 0.1	$s$
Fast	$10$	0.858	0.750	0.459	0.203	0.471	1.120	0.722
	$5$	0.877	0.768	0.467	0.204	0.480	1.125	0.724
	$4$	0.882	0.772	0.466	0.204	0.481	1.125	0.724
	$3$	0.887	0.775	0.465	0.201	0.481	1.125	0.724
	$2$	0.893	0.778	0.461	0.198	0.479	1.124	0.725
	$1$	0.898	0.778	0.454	0.193	0.475	1.122	0.726

Table 5: Results with varying risk functions. Fast-increasing risk functions with positive

\alpha

improve longer-horizon anticipation, i.e. ,

\geq 1

sec, while slow-increasing functions with negative

\alpha

perform better at shorter horizons, i.e. ,

<1

sec.

\alpha=3

achieves the best overall performance.

Panel of Figure 5 — Figure 5: Separation profile on CAP. BCE inflates pre-onset scores, whereas our model provides a smoother rise and improved accuracy near the accident, as shown zoomed-in.

	Model	$s_{\text{pre}}$ $\downarrow$	$s_{\text{post}}$ $\uparrow$	$s$ $\uparrow$
CAP	BCE	0.420	0.808	0.694
CAP	PRE-ACT	0.344	0.792	0.724
DADA	BCE	0.372	0.763	0.696
DADA	PRE-ACT	0.319	0.798	0.740

Table 6: Separation scores.

Appendix figures & tables3 assets

Supplementary material from the paper’s appendix.

Appendix

Dataset	$\lambda$	AUC ${}_{0.0s}^{\lambda}$	AUC ${}_{0.5s}^{\lambda}$	AUC ${}_{1.0s}^{\lambda}$	AUC ${}_{1.5s}^{\lambda}$	mAUC λ	mTTA λ
CAP [ 26 ]	0.01	0.721	0.500	0.106	0.025	0.210	0.817
	0.1	0.887	0.775	0.465	0.201	0.481	1.125
	1.0	0.976	0.945	0.852	0.701	0.833	1.820
DADA [ 39 ]	0.01	0.732	0.503	0.136	0.038	0.226	0.826
	0.1	0.861	0.751	0.398	0.170	0.440	1.069
	1.0	0.960	0.922	0.782	0.631	0.778	1.597

Appendix

Table 7: Performance across varying FPR. Despite drops at low FPR, our method maintains timely warnings, achieving

0.8

s mTTA at FPR

\leq 0.01

Dataset	Backbone	AUC ${}_{0.0s}^{0.1}$	AUC ${}_{0.5s}^{0.1}$	AUC ${}_{1.0s}^{0.1}$	AUC ${}_{1.5s}^{0.1}$	mAUC 0.1	mTTA 0.1
CAP [ 26 ]	DINOv2 [ 44 ]	0.837	0.573	0.272	0.136	0.327	0.721
	V-JEPAv2 [ 45 ]	0.881	0.778	0.426	0.176	0.460	1.120
	VideoMAE [ 40 ]	0.887	0.775	0.465	0.201	0.481	1.125
DADA [ 39 ]	DINOv2 [ 44 ]	0.674	0.358	0.179	0.117	0.218	0.850
	V-JEPAv2 [ 45 ]	0.596	0.516	0.314	0.177	0.336	1.011
	VideoMAE [ 40 ]	0.861	0.751	0.398	0.170	0.440	1.069

Appendix

Table 8: Varying backbone. Video models (V-JEPAv2 and VideoMAE) clearly outperform the image backbone (DINOv2), emphasizing spatio-temporal features; similar results across video encoders show backbone-agnostic behavior for our method.

Method	AP 0.5s	AP 1.0s	AP 1.5s	mAP	mAUC	mAUC 0.1	mTTA 0.1	$s$
PRE-ACT	0.919	0.911	0.854	0.895	0.895	0.510	1.285	0.665

Appendix

Table 9: Results using standard metrics on the Nexar [ 13 ] dataset.

Traffic Accident Anticipation

Traffic accident anticipation predicts imminent collisions or rising crash risk early enough to support driver warnings and proactive road-safety decisions.

See all 22 papers

RiskProp: Collision-Anchored Self-Supervised Risk Propagation for Early Accident Anticipation
Mar 28, 2026Yiyang Zou, Tianhao Zhao, Peilun Xiao +9Monotonicity Constraints Traffic Accident Anticipation
School of Computer Science, Wuhan University · Zhongguancun Academy, Beijing, China · Didi Chuxing +1
VAGNet: Vision-based Accident Anticipation with Global Features
Apr 10, 2026Vipooshan Vipulananthan, Charith D. ChitraranjanTraffic Accident Anticipation Autonomous Driving
Department of Computer Science and Engineering, University of Moratuwa, Katubedda 10400, Sri Lanka
FLaRA: Predicting Future Latent Representations for Accident Anticipation
Jun 12, 2026Lorenzo Caselli, Tomaso Trinci, Tommaso Bianconcini +4Autonomous Driving Traffic Accident Anticipation
Department of Information Engineering, University of Florence, Italy. · 2Verizon Connect, Florence, Italy.

Mar 28, 2026J/K move · Enter open · S save

RiskProp: Collision-Anchored Self-Supervised Risk Propagation for Early Accident Anticipation

Yiyang Zou, Tianhao Zhao, Peilun Xiao, Hongyu Jin+8

School of Computer Science, Wuhan University · Zhongguancun Academy, Beijing, China · Didi Chuxing+1

Monotonicity Constraints Traffic Accident Anticipation Self-Supervised Learning

Open paper AskPDF arXiv

Accident anticipation aims to predict impending collisions from dashcam videos and trigger early alerts. Existing methods rely on binary supervision with manually annotated "anomaly onset" frames, which are subjective and inconsistent, leading to inaccurate risk estimation. In contrast, we propose RiskProp, a novel collision-anchored self-supervised risk propagation paradigm for early accident anticipation, which removes the need for anomaly onset annotations and leverages only the reliably annotated collision frame. RiskProp models temporal risk evolution through two observation-driven losses: first, since future frames contain more definitive evidence of an impending accident, we introduce a future-frame regularization loss that uses the model's next-frame prediction as a soft target to supervise the current frame, enabling backward propagation of risk signals; second, inspired by the empirical trend of rising risk before accidents, we design an adaptive monotonic constraint to encourage a non-decreasing progression over time. Experiments on CAP and Nexar demonstrate that RiskProp achieves state-of-the-art performance and produces smoother, more discriminative risk curves, improving both early anticipation and interpretability.

Abstract

Traffic Accident Anticipation

Explore similar work

RiskProp: Collision-Anchored Self-Supervised Risk Propagation for Early Accident Anticipation

VAGNet: Vision-based Accident Anticipation with Global Features

FLaRA: Predicting Future Latent Representations for Accident Anticipation