A COMPARATIVE STUDY ON FACIAL EMOTION RECOGNITION (FER) USING CNN, CNN+LSTM & DENSENET
International Journal of Communication, Science, and Technology · 2025
DOI: 10.70965/ijcst.vol.1.issue2.8 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Communication, Science, and Technology, 2025. Verfasst von Subhabrata Bhattacharjee und 3 weiteren Autor:innen.
Abstract (Originalsprache)
The research evaluates three DL models for FER through their application to CNN, CNN-LSTM and DenseNet architectures.The models were tested on different facial expression datasets to evaluate their ability to detect human emotional cues.The CNN model functions as our initial spatial feature extraction tool while LSTM addition enables the network to detect temporal patterns in facial movements.The DenseNet network stands out by using feature reuse to maintain consistent gradient flow across its entire structure.Our experimental results demonstrate both the advantages and constraints of each model which leads to detailed understanding of their performance capabilities.The research outcomes enhance our comprehension of these architectures in emotion detection tasks while establishing essential foundations for future investigations into network optimization and real-time data processing.
Zitieren
APA
Bhattacharjee, S., Mitra, M., Chakraborty, P. K., & Chakladar, D. A. (2025). A COMPARATIVE STUDY ON FACIAL EMOTION RECOGNITION (FER) USING CNN, CNN+LSTM & DENSENET. International Journal of Communication, Science, and Technology. https://doi.org/10.70965/ijcst.vol.1.issue2.8
Vancouver
Bhattacharjee S, Mitra M, Chakraborty PK, Chakladar DA. A COMPARATIVE STUDY ON FACIAL EMOTION RECOGNITION (FER) USING CNN, CNN+LSTM & DENSENET. International Journal of Communication, Science, and Technology. 2025. doi:10.70965/ijcst.vol.1.issue2.8
BibTeX
@article{Bhattacharjee2025,
author = {Bhattacharjee, Subhabrata and Mitra, Mainak and Chakraborty, Prof. Koyel and Chakladar, Dr. Amit},
title = {A COMPARATIVE STUDY ON FACIAL EMOTION RECOGNITION (FER) USING CNN, CNN+LSTM & DENSENET},
journal = {International Journal of Communication, Science, and Technology},
year = {2025},
doi = {10.70965/ijcst.vol.1.issue2.8}
}
EUVIMED