中央研究院 資訊科學研究所

Holistic understanding of interactive scenarios is a vital component in the advancement of vision-based intelligent driving systems. In this talk, I will present two recent developments in the field of interactive traffic understanding. First, I will introduce a topology-aware traffic pattern description that allows for detailed road behavior analysis, alongside a slot attention-based recognition algorithm that supports weakly-supervised traffic pattern localization. Second, I will present a novel perspective on risk identification through vision-based analysis of driver behavior changes. We propose an risk identification algorithm inspired by the Situation Awareness model and causal inference. To conclude, I will shed light on the unresolved questions in the domain of visual interactive scenario understanding.

Yi-Ting Chen is an assistant professor in the Department of Computer Science at National Yang Ming Chiao Tung University. He received his Ph.D. degree from the Department of Electrical and Computer Engineering at Purdue University in 2015 and a B.S. degree in Electronics Engineering from National Chiao Tung University in 2009. He was a senior scientist at Honda Research Institute USA from 2015 to 2020. He was visiting researcher at Google ATAP in 2015 and a computer vision researcher at the University of California at Merced in 2014. His research interests lie in computer vision, machine learning, robotics, and behavior science, and their applications to intelligent systems. He is a recipient of Junior Faculty Award (豐達科技青年講座) from NYCU in 2021.