Article’s

Facial Emotion Recognition (FER) is an important application of Artificial Intelligence, Computer Vision, and Deep Learning that enables machines to identify human emotions through facial expressions. Human emotions such as happiness, sadness, anger, fear, surprise, disgust, and neutrality play a major role in communication and behavioral understanding. However, accurately classifying emotions remains challenging due to variations in facial appearance, image quality, lighting conditions, and similarities between emotional expressions. This project presents a hybrid Facial Emotion Recognition system using ensemble Deep Learning and Machine Learning techniques for accurate emotion classification. The system utilizes the FER-2013 Dataset consisting of grayscale facial images categorized into seven emotion classes. Multiple Convolutional Neural Network architectures including LeNet-5, ResNet-50, and VGG-16 are used for deep feature extraction. The extracted features are combined using ensemble learning and classified using a Support Vector Machine classifier. The proposed CNN-SVM hybrid architecture improves feature representation, reduces overfitting, and enhances classification robustness compared to traditional single-model systems. The developed FER system has potential applications in healthcare, surveillance systems, driver monitoring, gaming, robotics, and human-computer interaction.