Multi-Head Attention Affinity Diversity Sharing Network for Facial Expression Recognition

被引:0
|
作者
Zheng, Caixia [1 ,2 ]
Liu, Jiayu [2 ,3 ]
Zhao, Wei [4 ]
Ge, Yingying [1 ]
Chen, Wenhe [5 ,6 ]
机构
[1] Jilin Animat Inst, Sch Game, Changchun 130013, Peoples R China
[2] Northeast Normal Univ, Coll Informat Sci & Technol, Changchun 130117, Peoples R China
[3] Liaoning Sci & Technol Museum, Shenyang 110167, Peoples R China
[4] Jilin Police Coll, Dept Informat Secur Technol, Changchun 130117, Peoples R China
[5] Jiangsu Univ Technol, Sch Comp Engn, Changzhou 213001, Peoples R China
[6] Shanghai Huace Nav Technol Co Ltd, Shanghai 200000, Peoples R China
关键词
facial expression recognition; multi-head attention; discriminative feature representation; sparse affinity loss;
D O I
10.3390/electronics13224410
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Facial expressions exhibit inherent similarities, variability, and complexity. In real-world scenarios, challenges such as partial occlusions, illumination changes, and individual differences further complicate the task of facial expression recognition (FER). To further improve the accuracy of FER, a Multi-head Attention Affinity and Diversity Sharing Network (MAADS) is proposed in this paper. MAADS comprises a Feature Discrimination Network (FDN), an Attention Distraction Network (ADN), and a Shared Fusion Network (SFN). To be specific, FDN first integrates attention weights into the objective function to capture the most discriminative features by using the proposed sparse affinity loss. Then, ADN employs multiple parallel attention networks to maximize diversity within spatial attention units and channel attention units, which guides the network to focus on distinct, non-overlapping facial regions. Finally, SFN deconstructs facial features into generic parts and unique parts, which allows the network to learn the distinctions between these features without having to relearn complete features from scratch. To validate the effectiveness of the proposed method, extensive experiments were conducted on several widely used in-the-wild datasets including RAF-DB, AffectNet-7, AffectNet-8, FERPlus, and SFEW. MAADS achieves the accuracy of 92.93%, 67.14%, 64.55%, 91.58%, and 62.41% on these datasets, respectively. The experimental results indicate that MAADS not only outperforms current state-of-the-art methods in recognition accuracy but also has a relatively low computational complexity.
引用
收藏
页数:19
相关论文
共 50 条
  • [1] Distract Your Attention: Multi-Head Cross Attention Network for Facial Expression Recognition
    Wen, Zhengyao
    Lin, Wenzhong
    Wang, Tao
    Xu, Ge
    BIOMIMETICS, 2023, 8 (02)
  • [2] Local Multi-Head Channel Self-Attention for Facial Expression Recognition
    Pecoraro, Roberto
    Basile, Valerio
    Bono, Viviana
    INFORMATION, 2022, 13 (09)
  • [3] MPCSAN: multi-head parallel channel-spatial attention network for facial expression recognition in the wild
    Gong, Weijun
    Qian, Yurong
    Fan, Yingying
    NEURAL COMPUTING & APPLICATIONS, 2023, 35 (09): : 6529 - 6543
  • [4] MPCSAN: multi-head parallel channel-spatial attention network for facial expression recognition in the wild
    Weijun Gong
    Yurong Qian
    Yingying Fan
    Neural Computing and Applications, 2023, 35 : 6529 - 6543
  • [5] A facial depression recognition method based on hybrid multi-head cross attention network
    Li, Yutong
    Liu, Zhenyu
    Zhou, Li
    Yuan, Xiaoyan
    Shangguan, Zixuan
    Hu, Xiping
    Hu, Bin
    FRONTIERS IN NEUROSCIENCE, 2023, 17
  • [6] On the diversity of multi-head attention
    Li, Jian
    Wang, Xing
    Tu, Zhaopeng
    Lyu, Michael R.
    NEUROCOMPUTING, 2021, 454 : 14 - 24
  • [7] Deep Learning Based Mobilenet and Multi-Head Attention Model for Facial Expression Recognition
    Nouisser, Aicha
    Zouari, Ramzi
    Kherallah, Monji
    INTERNATIONAL ARAB JOURNAL OF INFORMATION TECHNOLOGY, 2023, 20 (3A) : 485 - 491
  • [8] Lightweight Facial Expression Recognition Based on Hybrid Multiscale and Multi-Head Collaborative Attention
    Zhang, Haitao
    Zhuang, Xufei
    Gao, Xudong
    Mao, Rui
    Ren, Qing-Dao-Er-Ji
    PATTERN RECOGNITION AND COMPUTER VISION, PRCV 2024, PT II, 2025, 15032 : 304 - 316
  • [9] Global multi-scale extraction and local mixed multi-head attention for facial expression recognition in the wild
    Fei, Zixiang
    Zhang, Bo
    Zhou, Wenju
    Li, Xia
    Zhang, Yukun
    Fei, Minrui
    NEUROCOMPUTING, 2025, 622
  • [10] Self Multi-Head Attention for Speaker Recognition
    India, Miquel
    Safari, Pooyan
    Hernando, Javier
    INTERSPEECH 2019, 2019, : 4305 - 4309