Deep Learning for Facial Emotion Recognition on the FER-2013 Dataset using the ResNet50v2 Model
Öz
Facial emotion recognition is a sophisticated approach that uses facial expression analysis to identify and understand human emotions. Its numerous applications in human-computer interaction, healthcare, and market research have recently received much attention. This technology aims to develop sophisticated algorithms and systems capable of accurately identifying and understanding an individual’s emotional state by analysing facial features. This study focuses on classifying human emotions using a deep learning model based on the ResNet50v2 architecture. This work presents a comprehensive facial expression recognition model utilising the FER-2013 dataset, which includes thousands of images annotated with seven distinct emotions: happy, angry, neutral, sad, disgust, fear, and surprise. Our approach involves several key steps, including significant image preprocessing to improve the input data quality, image transformation to increase the diversity and robustness of the model, and implementing a modified ResNet50v2 architecture to improve recognition accuracy. Our model achieved a 69% accuracy on the test data, demonstrating competitive performance compared to existing models applied to FER-2013. The results of this study highlight the great potential of deep learning methods in precisely identifying and deciphering human emotions from facial expressions, paving the way for more emotionally intelligent and responsive human-computer interaction systems.
Anahtar Kelimeler
Facial emotion recognition, Emotion classification, FER-2013, ResNet50v2
Kaynakça
- R. Buck, “Social and emotional functions in facial expression and communication: The readout hypothe -sis,” Biological psychology, vol. 38, no. 2-3, pp. 95-115, 1994.
- K. R. Scherer, A. Dieckmann, M. Unfried, H. Ellgring, and M. Mortillaro, “Investigating appraisal-driven facial expression and inference in emotion communication,” Emotion, vol. 21, no. 1, p. 73, 2021.
- T. Altameem and A. Altameem, “Facial expression recognition using human machine interaction and multi-modal visualization analysis for healthcare applications,” Image and Vision Computing, vol. 103, p. 104044, 2020.
- E. Arnaud, A. Dapogny, and K. Bailly, “Thin: Throwable information networks and application for facial expression recognition in the wild,” IEEE Transactions on Affective Computing, vol. 14, no. 3, pp. 2336-2348, 2022.
- M. Sajjad, M. Nasir, F. U. M. Ullah, K. Muhammad, A. K. Sangaiah, and S. W. Baik, “Raspberry Pi assisted facial expression recognition framework for smart security in law-enforcement services,” Information Sciences, vol. 479, pp. 416-431, 2019.
- X. Zhao, J. Zhu, B. Luo, and Y. Gao, “Survey on facial expression recognition: History, applications, and challenges,” IEEE multimedia, vol. 28, no. 4, pp. 38-44, 2021.
- M. Sajjad et al., “A comprehensive survey on deep facial expression recognition: challenges, applications, and future guidelines,” Alexandria Engineering Journal, vol. 68, pp. 817-840, 2023.
- S. Li and W. Deng, “Deep facial expression recognition: A survey,” IEEE transactions on affective compu-ting, vol. 13, no. 3, pp. 1195-1215, 2020. 32 Deep Learning for Facial Emotion Recognition... OMUJEST, 2025, Cilt 5, Sayı 2, Sayfa 19-33
- A. A. Pise, M. A. Alqahtani, P. Verma, P. K, D. A. Karras, and A. Halifa, “Methods for facial expression recogni-tion with applications in challenging situations,” Computational intelligence and neuroscience, vol. 2022, no. 1, p. 9261438, 2022.
- S. M. S. Abdullah and A. M. Abdulazeez, “Facial expression recognition based on deep learning convoluti-on neural network: A review,” Journal of Soft Computing and Data Mining, vol. 2, no. 1, pp. 53-65, 2021.