We update the driver attention restruction module and discussion of textual bias for traffic accident anticipation.

You can use the text generated by the PDVC and VTimeLLM-7B methods for model training. If you use the above text data for DADA-1000, please cite the PDVC and VTimeLLM-7B literatur.

More information please check out LOTVS-CAP

Reference

Lei-Lei Li, Jianwu Fang*, Jianru Xue, Cognitive Traffic Accident Anticipation. IEEE Intelligent Transportation Systems Magazine (ITSM), 2024.