ENHANCING FACIAL EXPRESSION AND ATTRIBUTES
RECOGNITION: AN EXPLORATION OF MULTI-TASK
LEARNING WITHIN LIGHTWEIGHT NEURAL NETWORKS
26.11.2023
International Scientific Journal "Science and Innovation". Series A. Volume 2 Issue 11
Abstract. Facial recognition, especially in the domains of expression and attribute
detection, has become pivotal in numerous applications. This study delves into the synergistic
integration of multi-task learning techniques with lightweight neural networks to address the dual
challenges of computational efficiency and robust performance. The research findings underscore
the potential of this combined approach, revealing a significant improvement in the recognition of
facial expressions and attributes. Furthermore, the proposed framework exhibits enhanced
efficiency, making it ideal for real-world applications that demand rapid and accurate facial
analysis.
Keywords: facial recognition, multi-task learning, Lightweight neural networks,
computational efficiency, robust performance, facial attributes, expression detection, real-world
applications.
References:
1. Agzamova M.Sh. Development of a software module implementing a proposed facial
biometric authentication algorithm and evaluation of solution effectiveness. SCIENCE AND
INNOVATION INTERNATIONAL SCIENTIFIC JOURNAL VOLUME 2 ISSUE 7 JULY
2023, pp. 51-57, https://doi.org/10.5281/zenodo.81507542.
2. Agzamova M.Sh., Irgasheva D.Y. Analysis of non-cryptographic methods for software
binding to facial biometric data of user identity. International Journal of Advance Scientific
Research, 3(07), 38–47. https://doi.org/10.37547/ijasr-03-07-08.
3. Agzamova M.Sh., Irgasheva D.Y. Analysis of facial authentication systems for neural
network modification of raw biometric data. Innovative Technologica: Methodical Research
Journal, 2(07), 16–28. https://doi.org/10.17605/OSF.IO/RZMFB.
4. Agzamova M.Sh., Irgasheva D.Y. A comprehensive review of the use of data mining
algorithms in facial recognition systems for payment systems. Bulletin of TUIT:
Management and Communication Technologies № 3(12)2023.
5. Chenqian Yan, Yuge Zhang, Quanlu Zhang, Yaming Yang, Xinyang Jiang, Yuqing Yang,
Baoyuan Wang. Privacy-preserving Online AutoML for Domain-Specific Face Detection.
URL: https://openaccess.thecvf.com/content/CVPR2022/papers/Yan_PrivacyPreserving_Online_AutoML_for_DomainSpecific_Face_Detection_CVPR_2022_paper.pdf
6. Yang Liu, Fei Wang, Jiankang Deng, Zhipeng Zhou, Baigui Sun, Hao Li. MogFace:
Towards a Deeper Appreciation on Face Detection. URL:
https://openaccess.thecvf.com/content/CVPR2022/papers/Liu_MogFace_Towards_a_Deep
er_Appreciation_on_Face_Detection_CVPR_2022_paper.pdf
7. Roberto Pecoraro, Valerio Basile, Viviana Bono, Sara Gallo. Local Multi-Head Channel
Self-Attention for Facial Expression Recognition. URL:
https://arxiv.org/pdf/2111.07224v2.pdf
8. Kai Wang, Xiaojiang Peng, Jianfei Yang, Debin Meng, Yu Qiao. Region Attention Networks
for Pose and Occlusion Robust Facial Expression Recognition. URL:
https://arxiv.org/pdf/1905.04075v2.pdf
9. Andrey V. Savchenko. Facial expression and attributes recognition based on multi-task
learning of lightweight neural networks. URL:
https://ieeexplore.ieee.org/abstract/document/9582508/authors#authors
10. Minchul Kim, Anil K. Jain, Xiaoming Liu. AdaFace: Quality Adaptive Margin for Face
Recognition. URL: https://arxiv.org/pdf/2204.00964.pd
![](images/google_scholar.png)
![](images/cyberleninka_logo.png)