Facial Emotion Recognition System

Abstract

Human feelings play a crucial function in powerful verbal exchange and choice-making. In the technology of artificial intelligence (AI) and human-computer interplay (HCI), enabling machines to apprehend, discover, and reply to human emotions has come to be increasingly essential. This research offers a real-time Facial Emotion Recognition (FER) system that detects and classifies seven key human feelings Angry, Disgust, Fear, Happy, Sad, Surprise, and Neutral from live video streams using a Convolutional Neural Network (CNN) model. The model changed into skilled at the FER2013 dataset, using facts augmentation, elegance balancing, and dropout techniques to improve accuracy. The machine uses OpenCV for real-time face detection, TensorFlow/Keras for CNN model education, and a Flask-based totally web application for an interactive person interface. The proposed technique demonstrates high actual-time overall performance, achieving as much as eighty five% validation accuracy, and offers a couple of actual-global packages, along with mental health monitoring, e-mastering, customer service enhancement, and emotion-aware AI assistants.