MICACL: Multi-Instance Category-Aware Contrastive Learning for Long-Tailed Dynamic Facial Expression Recognition

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Cui, Feng-Qi, Lin, Zhen, Rao, Xinlong, Tong, Anyang, Li, Shiyao, Wang, Fei, Chen, Changlin, Liu, Bin
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912799127502848
author Cui, Feng-Qi
Lin, Zhen
Rao, Xinlong
Tong, Anyang
Li, Shiyao
Wang, Fei
Chen, Changlin
Liu, Bin
author_facet Cui, Feng-Qi
Lin, Zhen
Rao, Xinlong
Tong, Anyang
Li, Shiyao
Wang, Fei
Chen, Changlin
Liu, Bin
contents Dynamic facial expression recognition (DFER) faces significant challenges due to long-tailed category distributions and complexity of spatio-temporal feature modeling. While existing deep learning-based methods have improved DFER performance, they often fail to address these issues, resulting in severe model induction bias. To overcome these limitations, we propose a novel multi-instance learning framework called MICACL, which integrates spatio-temporal dependency modeling and long-tailed contrastive learning optimization. Specifically, we design the Graph-Enhanced Instance Interaction Module (GEIIM) to capture intricate spatio-temporal between adjacent instances relationships through adaptive adjacency matrices and multiscale convolutions. To enhance instance-level feature aggregation, we develop the Weighted Instance Aggregation Network (WIAN), which dynamically assigns weights based on instance importance. Furthermore, we introduce a Multiscale Category-aware Contrastive Learning (MCCL) strategy to balance training between major and minor categories. Extensive experiments on in-the-wild datasets (i.e., DFEW and FERV39k) demonstrate that MICACL achieves state-of-the-art performance with superior robustness and generalization.
format Preprint
id arxiv_https___arxiv_org_abs_2509_04344
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle MICACL: Multi-Instance Category-Aware Contrastive Learning for Long-Tailed Dynamic Facial Expression Recognition
Cui, Feng-Qi
Lin, Zhen
Rao, Xinlong
Tong, Anyang
Li, Shiyao
Wang, Fei
Chen, Changlin
Liu, Bin
Computer Vision and Pattern Recognition
Dynamic facial expression recognition (DFER) faces significant challenges due to long-tailed category distributions and complexity of spatio-temporal feature modeling. While existing deep learning-based methods have improved DFER performance, they often fail to address these issues, resulting in severe model induction bias. To overcome these limitations, we propose a novel multi-instance learning framework called MICACL, which integrates spatio-temporal dependency modeling and long-tailed contrastive learning optimization. Specifically, we design the Graph-Enhanced Instance Interaction Module (GEIIM) to capture intricate spatio-temporal between adjacent instances relationships through adaptive adjacency matrices and multiscale convolutions. To enhance instance-level feature aggregation, we develop the Weighted Instance Aggregation Network (WIAN), which dynamically assigns weights based on instance importance. Furthermore, we introduce a Multiscale Category-aware Contrastive Learning (MCCL) strategy to balance training between major and minor categories. Extensive experiments on in-the-wild datasets (i.e., DFEW and FERV39k) demonstrate that MICACL achieves state-of-the-art performance with superior robustness and generalization.
title MICACL: Multi-Instance Category-Aware Contrastive Learning for Long-Tailed Dynamic Facial Expression Recognition
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2509.04344