Emotion detection from facial expression using image processing

Abstract

Abstract: Facial expression recognition is a powerful tool for communicating our emotions, understanding, and intent with each other. It is an intelligent human-computer interaction technology. Various studies have been conducted to classify facial expressions. Six fundamental universal emotions can be expressed through facial expressions: happiness, sadness, anger, fearful, surprised, and neutral. In this project, emotion detection can be implemented in real time with the help of a webcam. Our work proposed a CNN-based VGG16 architecture for emotion detection systems. A model would be trained by using the FER-2013 dataset. Then the images from the dataset are first pre-processed, which includes operations such as image scaling, changing the colour mode, and so on. Following that, a CNN model with multiple layers was created. After that, the model would be trained with the specified dataset, resulting in the .h5 file, which is a pre-trained model file. Instead of repeatedly training the model, the results can be predicted using this file.