Comparative Analysis of Transfer Learning Models for Facial Emotion Recognition
DOI:
https://doi.org/10.51611/iars.irj.v16i2.2026.287Keywords:
CNN, Facial Emotion Recognition, EfficientNetAbstract
Facial Emotion Recognition (FER) is an important component of human computer interaction, health monitoring, and affective computing systems. With the recent development of deep learning, transfer learning has gained popularity as an efficient technique for improving the classification accuracy and minimizing the training time, especially when working with a small dataset. In this paper, a comparative study of three popular pre-trained convolutional neural network models, namely ResNet, VGGNet, and EfficientNet, for facial emotion recognition is presented. The proposed models are fine-tuned on a standard facial emotion recognition dataset with seven basic classes of emotions, including anger, disgust, fear, happiness, sadness, surprise, and neutral. The performance analysis of the models is carried out using accuracy, precision, recall, F1-score, confusion matrix analysis, and computational complexity metrics like training time and model parameter size. Data augmentation strategies are used to improve the generalization capability of the models and avoid overfitting. The experimental outcomes show significant differences in terms of performance, complexity, and computational requirements among the three architectures. Although ResNet has a strong feature extraction ability because of the residual learning concept, VGGNet ensures a stable baseline performance with a less complex architecture. EfficientNet shows a competitive level of accuracy with a remarkably smaller number of parameters, which indicates its superior efficiency-performance trade-off. The experimental outcomes of this research work offer significant insights for researchers and practitioners to choose the most suitable transfer learning models for real-time emotion recognition systems, especially in resource-constrained and healthcare-based applications.
Downloads
References
Akcay, A. Atapour-Abarghouei, and T. P. Breckon, “Using Deep Convolutional Networks for Facial Emotion Recognition in the Wild,” IEEE International Conference on Image Processing (ICIP), 2016.
Dhall, R. Goecke, J. Joshi, M. Wagner, and T. Gedeon, “Emotion Recognition in the Wild: A Comparative Study,” IEEE Transactions on Affective Computing, 2020.
Fasel and J. Luettin, “Automatic facial expression analysis: A survey,” Pattern Recognition, vol. 36, no. 1, pp. 259–275, 2003.
Goodfellow et al., “Challenges in representation learning: A report on three machine learning contests,” Neural Networks, vol. 64, pp. 59–63, 2015.
He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016.
J. Pan and Q. Yang, “A survey on transfer learning,” IEEE Transactions on Knowledge and Data Engineering, vol. 22, no. 10, pp. 1345–1359, 2010.
Krizhevsky, I. Sutskever, and G. E. Hinton, “ImageNet classification with deep convolutional neural networks,” Advances in Neural Information Processing Systems (NeurIPS), 2012.
Li et al., “Deep Facial Expression Recognition: A Survey,” IEEE Transactions on Affective Computing, 2020.
Mahmoudi and M. Khalil, “A Survey of Transfer Learning in Facial Expression Recognition,” Journal of Visual Communication and Image Representation, 2021.
Mollahosseini, D. Chan, and M. H. Mahoor, “Going deeper in facial expression recognition using deep neural networks,” IEEE Winter Conference on Applications of Computer Vision (WACV), 2016.
Mollahosseini, D. Chan, and M. H. Mahoor, “Going Deeper in Facial Expression Recognition Using Deep Neural Networks,” IEEE Winter Conference on Applications of Computer Vision (WACV), 2016.
N. Levi and T. Hassner, “Emotion Recognition in the Wild via Convolutional Neural Networks and Metric Learning,” ACM International Conference on Multimodal Interaction, 2015.
Ojala, M. Pietikäinen, and T. Mäenpää, “Multiresolution gray-scale and rotation invariant texture classification with local binary patterns,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 24, no. 7, pp. 971–987, 2002.
S. N. Mishra and D. Jain, "Deep Learning-Based Robust 3D Face Recognition Under Diverse Pose and Facial Expression Variations," 2026 IEEE 18th International Conference on Computational Intelligence and Communication Networks (CICN), Manila, Philippines, 2026, pp. 347-352, doi: 10.1109/CICN70047.2026.11594274.
S. S. Sikarwar, H. Yennapusa, A. Agrawal, D. Jain, M. Malik and G. Sharma, "Using Smart Technology to Identify Pneumonia Accurately in Chest X-Rays," 2023 International Conference on Smart Devices (ICSD), Dehradun, India, 2024, pp. 1-5, doi: 10.1109/ICSD60021.2024.10751453.
Shaukat, M. Ahmad, and S. A. R. Zaidi, “Lightweight CNNs for Real-Time Facial Expression Recognition,” Pattern Recognition Letters, 2021.
Simonyan and A. Zisserman, “Very deep convolutional networks for large-scale image recognition,” International Conference on Learning Representations (ICLR), 2015.
Tan and Q. Le, “EfficientNet: Rethinking model scaling for convolutional neural networks,” International Conference on Machine Learning (ICML), 2019.
Tan and Q. Le, “EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks,” International Conference on Machine Learning (ICML), 2019.
W. Picard, Affective Computing. MIT Press, 1997.
Zhang and Z. Zhang, “Attention-Based ResNet for Facial Expression Recognition,” International Journal of Computer Vision and Robotics, 2019.
Zhang, “Transfer Learning for Facial Expression Recognition: A Comprehensive Analysis,” IEEE Access, 2019.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Megha Bansal, Neha Goel, Mitanshi Rastogi, Tanvi Jain
This work is licensed under a Creative Commons Attribution 4.0 International License.
Author(s) hold complete right on the content of this article. Copyright to the content are governed as per Copyright Policy of the Journal.
