Anticipating Traffic Accidents with Adaptive Loss and Large-scale Incident DB
This work addresses the problem of early traffic accident prediction for autonomous driving and safety systems, representing an incremental advance with specific performance gains.
The paper tackles traffic accident anticipation by introducing an adaptive loss function and a large-scale near-miss incident database, achieving state-of-the-art results with improvements of +6.6% mAP and 2.36 seconds earlier anticipation for risk anticipation, and +4.3% mAP and 0.70 seconds earlier for risk-factor anticipation.
In this paper, we propose a novel approach for traffic accident anticipation through (i) Adaptive Loss for Early Anticipation (AdaLEA) and (ii) a large-scale self-annotated incident database for anticipation. The proposed AdaLEA allows a model to gradually learn an earlier anticipation as training progresses. The loss function adaptively assigns penalty weights depending on how early the model can an- ticipate a traffic accident at each epoch. Additionally, we construct a Near-miss Incident DataBase for anticipation. This database contains an enormous number of traffic near- miss incident videos and annotations for detail evaluation of two tasks, risk anticipation and risk-factor anticipation. In our experimental results, we found our proposal achieved the highest scores for risk anticipation (+6.6% better on mean average precision (mAP) and 2.36 sec earlier than previous work on the average time-to-collision (ATTC)) and risk-factor anticipation (+4.3% better on mAP and 0.70 sec earlier than previous work on ATTC).