Constrained and directional ensemble attention for facial action unit detection

Open Webpage

Zhiwen Shao, Bikuan Chen, Yong Zhou, Xuehuai Shi, Canlin Li, Lizhuang Ma, Dit-Yan Yeung

Published: 2026, Last Modified: 30 Oct 2025Pattern Recognit. 2026EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•We propose to constrain self-attention to capture AU-related local information.•We propose to capture correlations among AUs by learning dual-directional attentions.•We introduce a multi-view ensemble method to merge the strengths of both models.•Our methods achieve state-of-the-art performance on challenging datasets.