Enhancing Facial Expression Recognition and Analysis with EfficientNetB7

Chenxin Huang

2024

Abstract

This paper introduces an advanced Facial Expression Recognition model utilizing EfficientNetB7 architecture. Through meticulous stages of data preprocessing, model training, testing, and evaluation, significant strides are made in accurately classifying various facial expressions. The model exhibits commendable performance, particularly excelling in identifying positive expressions. However, challenges persist in effectively distinguishing between neutral and sad expressions, warranting further investigation. Future enhancements could involve refining data preprocessing techniques, such as adversarial training and data synthesis, to bolster dataset diversity and robustness. Additionally, exploring more potent feature extraction methods, including ensemble learning and transfer learning, holds promise for augmenting recognition capabilities. To address nuances in neutral and sad expressions, integrating contextual cues or dynamic features into the model architecture is proposed. Moreover, enriching the model's understanding by incorporating auxiliary information like emotion dictionaries or sentiment labels offers a viable avenue for improvement. Overall, this study contributes insights and pathways for advancing Facial Expression Recognition systems towards greater accuracy and applicability in real-world scenarios.

Paper Citation

in Harvard Style

Huang C. (2024). Enhancing Facial Expression Recognition and Analysis with EfficientNetB7. In Proceedings of the 1st International Conference on Engineering Management, Information Technology and Intelligence - Volume 1: EMITI; ISBN 978-989-758-713-9, SciTePress, pages 490-495. DOI: 10.5220/0012956900004508

in Bibtex Style

@conference{emiti24,
author={Chenxin Huang},
title={Enhancing Facial Expression Recognition and Analysis with EfficientNetB7},
booktitle={Proceedings of the 1st International Conference on Engineering Management, Information Technology and Intelligence - Volume 1: EMITI},
year={2024},
pages={490-495},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0012956900004508},
isbn={978-989-758-713-9},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 1st International Conference on Engineering Management, Information Technology and Intelligence - Volume 1: EMITI
TI - Enhancing Facial Expression Recognition and Analysis with EfficientNetB7
SN - 978-989-758-713-9
AU - Huang C.
PY - 2024
SP - 490
EP - 495
DO - 10.5220/0012956900004508
PB - SciTePress