核心概念
Facial AU detection benefits from contrastive learning for person-independent representations.
要約
The content discusses the challenges of AU detection due to data scarcity and proposes a contrastive learning approach to learn AU representations from unlabeled facial videos. The method aims to encode discriminative AU representations within video clips and across different identities showing similar AUs. Experimental results show the effectiveness of the proposed method in improving AU detection performance.
- Introduction to Facial Action Unit (AU) detection and the challenges of data scarcity.
- Proposal of contrastive learning for AU representation learning from unlabeled facial videos.
- Explanation of the method involving intra-video contrastive learning and cross-identity reconstruction.
- Details on the training objectives and implementation of the proposed method.
- Evaluation of the method on three popular AU datasets and comparison with state-of-the-art methods.
- Results showing the effectiveness of the proposed method in improving AU detection performance.
統計
"Experimental results on three public AU datasets demonstrate that the learned AU representation is discriminative for AU detection."
"The proposed method outperforms other contrastive learning methods and significantly closes the performance gap between self-supervised and supervised AU detection approaches."
引用
"We propose to contrastively learn the AU representation within a video clip and devise a cross-identity reconstruction mechanism to learn the person-independent representations."
"Our method outperforms other contrastive learning methods and significantly closes the performance gap between the self-supervised and supervised AU detection approaches."