Convolutional Neural Network-Based Contemporaneous Human Facial Expression Identification

In this paper, we discuss the use of facial expression in tandem with feature extraction and neural network to recognize distinct facial emotions like happy, sad, angry, surprised, and neutral. We look at the limitations of existing emotion identification algorithms using convolutional neural networks on CK+ and modified FER13 datasets and achieved decent prediction. Expression recognition utilizing a brain activity system is more difficult and time-consuming than facial emotion recognition using CNN. Alongside with recognizing the user's emotion, the paper also aims to analyze and generate an emotion log file comprising all of the emotions recorded with a time stamp. We have achieved 97.87% accuracy on FER13 dataset, and for CK+ dataset in testing, we got 94.93% of accuracy. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.