Facial Expression-Based Emotion Detection Using Deep Learning
International Journal of Science and Research (IJSR)

International Journal of Science and Research (IJSR)
Call for Papers | Fully Refereed | Open Access | Double Blind Peer Reviewed

ISSN: 2319-7064


Downloads: 1 | Views: 43 | Weekly Hits: ⮙1 | Monthly Hits: ⮙1

Research Paper | Computer Science and Information Technology | United States of America | Volume 13 Issue 8, August 2024 | Popularity: 4.9 / 10


     

Facial Expression-Based Emotion Detection Using Deep Learning

Omkar Reddy Polu


Abstract: Affective computing, human - computer interacting (HCI) and behavioral analysis research areas has its own importance as the facial expression-based emotion detecting. Real time performance, complex occlusions and cross-cultural expressions are the difficulties that traditional methods cannot achieve. However, in this study, an advancing deep learning framework based on multi scale feature extraction, residual connections and attention mechanism is put forward to achieve the excellent classification accuracy in emotion classification. We base our model on an upgraded Convolutional Neural Network (CNN) augmented with self-attention transformer encoder using which we will be capturing the spatial and temporal dependencies. In this work, we introduce the architecture with a dual branch: a CNN branch for local feature extraction, as well as a Vision Transformer (ViT) branch for global context modeling. A cross - modal attention mechanism is also used to fuse features with a feature fusion strategy, thereby increasing robustness to lighting, pose and occlusion variations as well. To reduce the domain gap, rare expressions of the dataset are augmented using generative adversarial networks (GANs), which help the model generalize better. We also demonstrate on benchmark datasets such as FER 2013, CK+, and AffectNet, extensive improvement in accuracy against all state of art methods, and reach an 92.5%. In the existing model, we propose to apply quantization techniques and TensorRT to ensure it optimizes for real time deployment on edge devices. The work presented in this research is of great value to development of advanced emotion recognition systems in healthcare, security, and human robot interaction.


Keywords: Facial Expression Recognition, Deep Learning, Convolutional Neural Networks (CNN), Vision Transformer (ViT), Attention Mechanisms, Feature Fusion, Generative Adversarial Networks (GANs), Emotion Detection, Human - Computer Interaction (HCI), Real - Time Deployment


Edition: Volume 13 Issue 8, August 2024


Pages: 1994 - 1999


DOI: https://www.doi.org/10.21275/SR24049114053


Please Disable the Pop-Up Blocker of Web Browser

Verification Code will appear in 2 Seconds ... Wait



Text copied to Clipboard!
Omkar Reddy Polu, "Facial Expression-Based Emotion Detection Using Deep Learning", International Journal of Science and Research (IJSR), Volume 13 Issue 8, August 2024, pp. 1994-1999, https://www.ijsr.net/getabstract.php?paperid=SR24049114053, DOI: https://www.doi.org/10.21275/SR24049114053

Top