PAND: Precise Action Recognition on Naturalistic Driving

Hangyue Zhao, Yuchao Xiao, Yanyun Zhao; Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, 2022, pp. 3291-3299

Temporal action localization for untrimmed videos is a difficult problem in computer vision. It is challenge to infer the start and end of activity instances on small-scale datasets covering multi-view information accurately. In this paper, we propose an effective activity temporal localization and classification method to lo-calize the temporal boundaries and predict the class la-bel of activities for naturalistic driving. Our approach includes (i) a distraction behavior recognition and lo-calization method in naturalistic driving videos on small-scale data sets, (ii) a strategy that uses mul-ti-branch network to make full use of information from different channels, (iii)a post-processing method for se-lecting and correcting temporal range to ensure that our system finds accurate boundaries. In addition, the frame-level object detection information is also utilized. Extensive experiments prove the effectiveness of our method and we rank the 6th on the Test-A2 of the 6th AI City Challenge Track 3.