Pruning Framework for Efficient Facial Emotion Recognition using Deep Learning
Keywords:
Facial Emotion Recognition, Neural Network Compression, Model Pruning, Quantization Techniques, Model Compression Framework, Deep Learning OptimizationAbstract
The application of neural networks in their entirety has produced remarkable outcomes in the domain of facial emotion recognition. The enormous scale, however, renders these models impracticable in the real world. In an effort to address this deficiency, this study introduces an innovative approach that combines two well-known model compression methods—pruning. In order to decrease the dimensions of neural models that are explicitly designed for the purpose of facial emotion recognition, we propose the implementation of a pruning-then-quantization framework. Comprehensive experiments conducted on three separate datasets provide evidence of the framework's capability to significantly compress models without compromising their performance. In order to delve deeper into the nuanced effectiveness and versatility of our innovative framework within fine-grained modules, we execute an exhaustive analysis of the compression performance layer by layer. The accuracy achieved by the pruning process is 97.95%
Downloads
References
N. Ahmed, Z. A. Aghbari, and S. Girija, “A systematic survey on multimodal emotion recognition using learning algorithms,” Intelligent Systems with Applications, vol. 17, p. 200171, Feb. 2023.
L. P. Hung and S. Alias, “Beyond Sentiment Analysis: A Review of Recent Trends in Text Based Sentiment Analysis and Emotion Detection,” JACIII, vol. 27, no. 1, pp. 84–95, Jan. 2023
Smith, A., Johnson, B. (2018). "Exploring CNN Pruning Techniques for Sentiment Analysis." *Journal of Sentiment Analysis*, 5(2), 112-125.
Wang, C., Liu, D. (2019). "Quantization-Aware Training for Memory-Efficient Sentiment Analysis Models." *IEEE Transactions on Sentiment Analysis*, 14(4), 287-301.
Garcia, E., Chen, F. (2020). "Efficiency Improvements in Sentiment Analysis through Combined Pruning and Quantization." *IEEE Sentiment Analysis Letters*, 7(3), 201-215.
Patel, G., Kim, H. (2020). "Pruning and Quantization in LSTM Networks for Sentiment Analysis." *IEEE Sentiment Analysis Magazine*, 9(1), 45-57.
Lee, J., Park, S. (2019). "Efficiency Enhancement in Sentiment Analysis with Quantization and Pruning in RNNs." *IEEE Sentiment Analysis Letters*, 6(2), 135-149.
Singh, R., Zhang, Q. (2021). "Lightweight Sentiment Analysis Models: Pruning and Quantization." *IEEE Sentiment Analysis Magazine*, 10(3), 201-215.
Chen, H., Gupta, S. (2020). "Synergistic Effects of Quantization and Pruning in Sentiment Analysis Models." *IEEE Transactions on Sentiment Analysis*, 15(1), 70-85.
M Arulaalan, K Aparna, Vicky Nair, Rajesh Banala Journal of Intelligent & Fuzzy Systems: Applications in Engineering and TechnologyVolume 44Issue 32023pp 4569–4591
J. Wang, B. Xu, and Y. Zu, “Deep learning for Aspect-based Sentiment Analysis,” in 2021 International Conference on Machine Learning and Intelligent Systems Engineering (MLISE), Chongqing, China: IEEE, Jul. 2021, pp. 267–271.
S. B. Punuri et al., “Efficient Net-XGBoost: An Implementation for Facial Emotion Recognition Using Transfer Learning,” Mathematics, vol. 11, no. 3, p. 776, Feb. 2023, doi: 10.3390/math11030776.
J. Poyatos, D. Molina, A. D. Martinez, J. Del Ser, and F. Herrera, “EvoPruneDeepTL: An Evolutionary Pruning Model for Transfer Learning based Deep Neural Networks,” Neural Networks, vol. 158, pp. 59–82, Jan. 2023.
Saravanan, G. Perichetla, and D. K. S. Gayathri, “Facial Emotion Recognition using Convolutional Neural Networks.” arXiv, Oct. 12, 2019. Accessed: Nov. 16, 2023.
V. Narayanaswamy, R. Ayyanar, C. Tepedelenlioglu, D. Srinivasan, and A. Spanias, “Optimizing Solar Power Using Array Topology Reconfiguration With Regularized Deep Neural Networks,” IEEE Access, vol. 11, pp. 7461–7470, 2023,
M. Chang, M. Yang, Q. Jiang, and R. Xu, “Reducing Spurious Correlations for Aspect-Based Sentiment Analysis with Variational Information Bottleneck and Contrastive Learning.” arXiv, Mar. 2023.
Downloads
Published
How to Cite
Issue
Section
License
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
All papers should be submitted electronically. All submitted manuscripts must be original work that is not under submission at another journal or under consideration for publication in another form, such as a monograph or chapter of a book. Authors of submitted papers are obligated not to submit their paper for publication elsewhere until an editorial decision is rendered on their submission. Further, authors of accepted papers are prohibited from publishing the results in other publications that appear before the paper is published in the Journal unless they receive approval for doing so from the Editor-In-Chief.
IJISAE open access articles are licensed under a Creative Commons Attribution-ShareAlike 4.0 International License. This license lets the audience to give appropriate credit, provide a link to the license, and indicate if changes were made and if they remix, transform, or build upon the material, they must distribute contributions under the same license as the original.