Proceedings of the International Conference on Sustainable Science and Technology for Tomorrow (SciTech 2024)

<Previous Article In Volume
Next Article In Volume>

Optimizing CNN Models for Facial Expression Recognition: A Comparative Study of Fine-Tuning Impact

Authors
Hrishiraj Sawan1, Riya Deka2, Sagar Saikia3, *, Sunil Prasad4
1Department of Computer Science and Engineering, Indian Institute of Information Technology, Guwahati, 781015, India
2Department of Computer Science and Information Technology, Cotton University, Guwahati, 781001, India
3Department of Computer Science and Engineering, National Institute of Technology, Meghalaya, 793003, India
4Department of Social Work, Indira Gandhi University, Rewari, Haryana, 122502, India
*Corresponding author. Email: p19cs010@nitm.ac.in
Corresponding Author
Sagar Saikia
Available Online 23 October 2025.
DOI
10.2991/978-94-6463-876-9_14How to use a DOI?
Keywords
CNN models; Facial Expression Recognition (FER); CNN architecture comparison; InceptionResNetV2 performance; deep learning optimization; emotion detection models; fine-tuning impact; MobileNetV2 efficiency; GoogleNet; transfer learning in FER
Abstract

InceptionResNetV2, MobileNetV2, and GoogleNet are three cutting-edge convolutional neural network (CNN) models that are compared in this paper’s study of facial expression recognition prediction. The models were tested on a dataset collected from 25 participants with emotions like fear, sadness, disgust, and happiness. After fine-tuning, InceptionResNetV2 achieved 98.50% accuracy, GoogleNet reached 97.94%, and MobileNetV2 improved significantly to 84.43%. This study highlights the trade-offs between model complexity and performance, and the impact of fine-tuning on optimization.

This comparative analysis provides insights into the most effective CNN architecture for facial expression recognition and also subsequently classifying it, offering guidance for researchers and practitioners in the field of computer vision.

Copyright
© 2025 The Author(s)
Open Access
Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

<Previous Article In Volume
Next Article In Volume>
Volume Title
Proceedings of the International Conference on Sustainable Science and Technology for Tomorrow (SciTech 2024)
Series
Atlantis Advances in Applied Sciences
Publication Date
23 October 2025
ISBN
978-94-6463-876-9
ISSN
3091-4442
DOI
10.2991/978-94-6463-876-9_14How to use a DOI?
Copyright
© 2025 The Author(s)
Open Access
Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Cite this article

risenwbib
TY  - CONF
AU  - Hrishiraj Sawan
AU  - Riya Deka
AU  - Sagar Saikia
AU  - Sunil Prasad
PY  - 2025
DA  - 2025/10/23
TI  - Optimizing CNN Models for Facial Expression Recognition: A Comparative Study of Fine-Tuning Impact
BT  - Proceedings of the International Conference on Sustainable Science and Technology for Tomorrow (SciTech 2024)
PB  - Atlantis Press
SP  - 163
EP  - 179
SN  - 3091-4442
UR  - https://doi.org/10.2991/978-94-6463-876-9_14
DO  - 10.2991/978-94-6463-876-9_14
ID  - Sawan2025
ER  -