Publication

Abstract : The ability to recognize facial emotions through computer vision has become a challenging yet crucial task in the field of image classification. This paper introduces a method for detecting facial emotions, focusing on the classification of seven distinct emotions which are neutral, sadness, anger, disgust, fear, happiness, and surprise. Using recent advancements in computer vision and machine learning, the proposed approach employs three distinct models: Convolutional Neural Network (CNN), ResNet50, and InceptionV3. Facial expression recognition poses unique challenges in image classification, and the adoption of deep learning has gained prominence for its effectiveness in solving this complex problem. The key challenge addressed in this research is the design of architectures that balance both simplicity and effectiveness. A simple architecture ensures quick training and easy implementation, while an effective one yields high accuracy on the test data. This research contributes to the ongoing efforts in advancing computer vision applications related to human emotions and highlights the potential for practical implementation in areas like affective computing and human computer interaction. The findings offer valuable insights into designing effective and efficient deep-learning architectures for facial emotion recognition with a multiclass approach.