Enhancing Facial Emotion Recognition Through Deep Learning: Integrating CNN and RNN-LSTM Models

Jianhui Xu

2024

Abstract

This article investigates the application of deep learning techniques in Facial Emotion Recognition (FER) to advance psychological research and practical applications. Given its increasing relevance for improving human-computer interaction, mental health assessment, and accessibility for individuals with disabilities, FER is a field of growing importance. The proposed method combines Convolutional Neural Networks (CNN) for extracting spatial features from facial images with Recurrent Neural Networks (RNNs) and Long Short-Term Memory (LSTM) units for analyzing temporal evolution, particularly in video sequences. CNNs are employed to discern subtle variations in facial expressions, while RNN-LSTM models capture the progression of emotions over time. Experiments conducted on the FER-2013 and AffectNet datasets demonstrate that the CNN model outperforms other models, achieving accuracy levels that exceed those of human recognition on the FER-2013 dataset. This integration of CNN and RNN-LSTM models holds significant promise for enhancing the accuracy and efficiency of FER systems. Future research will focus on mitigating cultural biases, optimizing real-time application performance, and addressing privacy and ethical considerations in FER technology deployment.

Paper Citation

in Harvard Style

Xu J. (2024). Enhancing Facial Emotion Recognition Through Deep Learning: Integrating CNN and RNN-LSTM Models. In Proceedings of the 2nd International Conference on Data Analysis and Machine Learning - Volume 1: DAML; ISBN 978-989-758-754-2, SciTePress, pages 131-136. DOI: 10.5220/0013510700004619

in Bibtex Style

@conference{daml24,
author={Jianhui Xu},
title={Enhancing Facial Emotion Recognition Through Deep Learning: Integrating CNN and RNN-LSTM Models},
booktitle={Proceedings of the 2nd International Conference on Data Analysis and Machine Learning - Volume 1: DAML},
year={2024},
pages={131-136},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013510700004619},
isbn={978-989-758-754-2},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 2nd International Conference on Data Analysis and Machine Learning - Volume 1: DAML
TI - Enhancing Facial Emotion Recognition Through Deep Learning: Integrating CNN and RNN-LSTM Models
SN - 978-989-758-754-2
AU - Xu J.
PY - 2024
SP - 131
EP - 136
DO - 10.5220/0013510700004619
PB - SciTePress