Explore the Facial Expression Recognition in Different Complex Environment

Haotian Lu; Jiayi Xu; Haoteng Zheng

doi:10.54254/2755-2721/2025.PO24884

Applied and Computational EngineeringOpen access

Explore the Facial Expression Recognition in Different Complex Environment

Research Article

Open Access

Explore the Facial Expression Recognition in Different Complex Environment

Haotian Lu ^1* Jiayi Xu ², Haoteng Zheng ³

¹ School of Shanghai Liaoyuan Bilingual, Shanghai, 200336, China

² Shanghai Nanyang High School, Shanghai, 200032, China

³ School of Computer Science and Engineering, Central South University, Changsha, Hunan, 410083, China

^*Corresponding author: 8208220515@csu.edu.cn

Published on 10 July 2025

ACE Vol.174

ISSN (Print): 2755-273X

ISSN (Online): 2755-2721

ISBN (Print): 978-1-80590-235-5

ISBN (Online): 978-1-80590-236-2

Download Cover

Abstract

Facial Expression Recognition (FER) holds extensive application value in fields such as human-computer interaction, intelligent security, and affective computing. However, it also faces challenges from complex conditions like illumination variations, occlusions, and head pose deviations, which cause significant performance degradation in traditional methods. This paper systematically analyzes recent advances in FER technologies addressing these issues and explores the application and innovation of deep learning methods in this domain. The analysis focuses on the following research directions: (1) Enhancement of recognition rate and robustness under illumination variations. Improved models incorporating transfer learning, activation function optimization, and anti-aliasing techniques have significantly boosted recognition accuracy. Approaches such as self-supervised learning and multi-feature aggregation demonstrate strong robustness and adaptability. (2) Feature compensation in occluded scenarios. Effective representation and facial feature-based methods are employed to restore occluded regions, improving recognition efficiency while fully leveraging the spatial structure of occlusions. This enables better representation and reconstruction of original samples. (3) Multi-pose facial expression recognition. By utilizing 3D deformable models for pose estimation and designing pose-invariant feature extraction networks, the challenges of expression recognition under large-angle head rotations are effectively addressed. This paper outlines the limitations of existing approaches and provides an outlook on potential future application scenarios and directions for technological breakthroughs.

Keywords:

Facial Expression Recognition, Unconstrained Environments, Deep Learning, Self-Supervised Learning, Computer Vision

View PDF

References

[1]. H. Ali, W. K. Wan Ahmad, H. Muthusamy, and M. Elshaikh, "Illumination effects on facial expression recognition using empirical mode decomposition, " in Proc. ICAROB, vol. 29, pp. 653-660, Feb. 2024.

[2]. K. P. Rao and M. V. P. C. S. Rao, "Illumination invariant facial expression recognition using convolutional neural networks, " Int. J. Recent Technol. Eng., vol. 8, no. 4, pp. 6140-6144, Nov. 2019.

[3]. S. Kuruvayil and S. Palaniswamy, "Emotion recognition from facial images with simultaneous occlusion, pose and illumination variations using meta-learning, " Journal of King Saud University - Computer and Information Sciences, vol. 34, no. 9, pp. 7271-7282, 2022.

[4]. A. Ruiz-Garcia, V. Palade, M. Elshaw and I. Almakky, "Deep Learning for Illumination Invariant Facial Expression Recognition, " 2018 International Joint Conference on Neural Networks (IJCNN), pp. 1-6, 2018.

[5]. Karu Prasada Rao and M. V. P. C. S. Rao, "Illumination Invariant Facial Expression Recognition using Convolutional Neural Networks." International Journal of Recent Technology and Engineering (IJRTE) 8.4 (2019): 6140-6144.

[6]. R.A. Elsheikh et al., "Improved facial emotion recognition model based on a novel deep convolutional structure." Scientific Reports 14 (2024): 29050.

[7]. Sukhavasi SB et al., "Deep Neural Network Approach for Pose, Illumination, and Occlusion Invariant Driver Emotion Detection." International Journal of Environmental Research and Public Health 19.4 (2022): 2352.

[8]. Huanjie Tao and Qianyue Duan, "Hierarchical attention network with progressive feature fusion for facial expression recognition." Neural Networks 170 (2024): 337-348.

[9]. Cui Xinyu, He Chong, Zhao Hongke et al. "Facial expression recognition by integrating ViT and contrastive learning." Journal of Image and Graphics, 29.1 (2024): 123-133.

[10]. Licai Sun et al., "MAE-DFER: Efficient Masked Autoencoder for Self-supervised Dynamic Facial Expression Recognition." Proceedings of the 31st ACM International Conference on Multimedia (2023).

[11]. Yang M et al., "Fisher discrimination dictionary learning for sparse representation." IEEE International Conference on Computer Vision (2011): 543-547.

[12]. Wang Huixing, Huang Bo, Gao Yongbin et al. "A review of some methods for face recognition with partial occlusion." Journal of Wuhan University (Science Edition) 66.5 (2020): 451-461.

[13]. OU W H，LUAN X，GOU J PJ N al. Robust discriminative nonnegative dictionary learning for occluded face recognition ［J］. Pattern Recognition Letters，2018，107：41-49. DOI：10.1016/j.patrec.2017.07.006.

[14]. Lianghai Jin et al., "Quaternion Deformable Local Binary Pattern and Pose-Correction Facial Decomposition for Color Facial Expression Recognition in the Wild." IEEE Transactions on Computational Social Systems 11.2 (2024).

[15]. Bulat A and Tzimiropoulos G, "How far are we from solving the 2D & 3D face alignment problem? (And a dataset of 230, 000 3D facial landmarks)." IEEE International Conference on Computer Vision (2017): 1021-1030.

[16]. Xiao Li et al., "Accurate Head Pose Estimation Using Image Rectification and a Lightweight Convolutional Neural Network." IEEE Transactions on Multimedia 25 (2023).

[17]. Yaoxing Wang et al., "HeadDiff: Exploring Rotation Uncertainty With Diffusion Models for Head Pose Estimation." IEEE Transactions on Image Processing 33 (2024).