Group commonality graph: Multimodal pedestrian trajectory prediction via deep group features

Open Webpage

Di Zhou, Ying Gao, Hui Li, Xiaoya Liu, Qinghua Lin

Published: 01 Jan 2025, Last Modified: 06 Nov 2025Pattern Recognit. Lett. 2025EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•Our model considers the effect of individual differences on group behavior.•Design a bi-layer threshold structure to reduce misjudgment and improve stability.•Propose a new multimodal predictive evaluation metric incorporating map obstacle information.•The framework is successfully applied to multimodal trajectory prediction benchmark dataset.