Deep Learning-Based Human Emotion Detection Framework Using Facial Expressions

Abstract

Automatic recognition of facial expression is an emerging study in the recognition of emotions. Emotion plays a significant role in understanding people and is usually related to sound decisions, behaviors, human activities, and intellect. The scientific community needs accurate and deployable technologies to understand human beings’ emotional states to establish practical and emotional interactions between human beings and machines. In the paper, a deep learning-based human emotion detection framework (DL-HEDF) has been proposed to evaluate the probability of digital representation, identification, and estimation of feelings. The proposed DL-HEDF analyzes the impact of emotional models on multimodal identification. The paper introduces emerging works that use existing methods like convolutional neural networks (CNN) for human emotion identification based on language, sound, image, video, and physiological signals. The proposed emphasis on the province study illustrates the shape and display of sample size emotional stimulation. While the findings obtained are not a province, the evidence collected indicates that deep learning could be sufficient to classify face emotion. Deep learning can enhance interaction with people because it allows computers to acquire perception by learning characteristics. And by perception, robots can offer better responses, enhancing the user experience dramatically. Six basic emotional levels have been successfully classified. The suggested way of recognizing emotions has then proven effective. The output results are obtained as an analysis of the ratio of the facial expression of 87.16%, accuracy evaluation ratio being 88.7%, improving facial recognition ratio is 84.5%, and the expression intensity ratio is 82.2%. The emotional simulation ratio is 93.0%.