---
title: "Towards Regularized Mixture of Predictions for Class-Imbalanced Semi-Supervised Facial Expression Recognition"
authors:
  - "Hangyu Li"
  - "Yixin Zhang"
  - "Jiangchao Yao"
  - "Nannan Wang"
  - "Bo Han"
doi: "10.24963/ijcai.2025/154"
url: "https://ijcai.org/proceedings/2025/154"
pdf: "https://ijcai.org/proceedings/2025/0154.pdf"
venue: "Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence"
year: 2025
pages: "1377-1385"
track: "Main Track"
keywords:
  - "Computer Vision: CV: Transfer, low-shot, semi- and un- supervised learning"
  - "Computer Vision: CV: Biometrics, face, gesture and pose recognition"
  - "Humans and AI: HAI: Human-computer interaction"
---

# Towards Regularized Mixture of Predictions for Class-Imbalanced Semi-Supervised Facial Expression Recognition

Hangyu Li, Yixin Zhang, Jiangchao Yao, Nannan Wang, Bo Han

*Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence*, pages 1377-1385. DOI: [10.24963/ijcai.2025/154](https://doi.org/10.24963/ijcai.2025/154)

## Abstract

Semi-supervised facial expression recognition (SSFER) effectively assigns pseudo-labels to confident unlabeled samples when only limited emotional annotations are available. Existing SSFER methods are typically built upon an assumption of the class-balanced distribution. However, they are far from real-world applications due to biased pseudo-labels caused by class imbalance. To alleviate this issue, we propose Regularized Mixture of Predictions (ReMoP), a simple yet effective method to generate high-quality pseudo-labels for imbalanced samples. Specifically, we first integrate feature similarity into the linear prediction to learn a mixture of predictions. Furthermore, we introduce a class regularization term that constrains the feature geometry to mitigate imbalance bias. Being practically simple, our method can be integrated with existing semi-supervised learning and SSFER methods to tackle the challenge associated with class-imbalanced SSFER effectively. Extensive experiments on four facial expression datasets demonstrate the effectiveness of the proposed method across various imbalanced conditions. The source code is made publicly available at https://github.com/hangyu94/ReMoP.