Review
Jul 2026
Vision-language and generative models in traffic video safety analysis: a computational framework and research agenda.
A unified taxonomy is proposed that maps model families to three application levels across diverse deployment environments-from cloud to onboard systems: network-scale monitoring, event-level crash and near-miss understanding, and generative or counterfactual scenario analysis.
Chen-Zhu Wang, Mohamed A. Abdel-Aty, Said M. Easa et al.
· Accident Analysis and Preven... · 0 citations