Recognition of human emotions in real time

DOI: 10.31673/2412-9070.2023.065457

  • Варфоломеєва О. Г. (Varfolomeieva O. G.) State University of Information and Communication Technologies, Kyiv
  • Отрох С. І. (Otrokh S. I.) National Technical University of Ukraine “Igor Sikorsky Kyiv Polytechnic Institute”, Kyiv
  • Сушильников К. Д. (Sushylnykov K. D.) National Technical University of Ukraine “Igor Sikorsky Kyiv Polytechnic Institute”, Kyiv
  • Шалигін М. О. (Shalyhin M. O.) National Technical University of Ukraine “Igor Sikorsky Kyiv Polytechnic Institute”, Kyiv

Abstract

This article emphasizes the importance and relevance of using neural networks to recognize human emotions in real time. He explores the possibilities of using convolutional neural networks for efficient analysis and classification of emotions based on facial expressions. The article describes the process of developing an application using JavaScript and TensorFlow and highlights potential optimization opportunities to improve the speed and accuracy of emotion recognition. The solution proposed in this paper offers a highly effective approach to emotion recognition, applicable in a variety of fields ranging from medicine and social research to marketing and entertainment. This highlights the importance of advanced technologies that improve human-computer interaction and reflects the importance of using innovative methods to improve the quality of human life. The convolutional neural network (CNN) algorithm used in this study uses a deep learning approach for facial expression recognition. The architecture of the CNN model includes several convolutional layers, each of which is responsible for identifying specific features of face images, such as edges, textures, and patterns. These layers are followed by pooling layers to reduce dimensionality, allowing the network to focus on the most meaningful features. In addition, the use of rectified linear units (ReLU) as activation functions improves the model’s ability to capture complex relationships in the data. In addition, the paper examines a deep learning facial expression recognition (FER) model based on convolutional neural networks, demonstrating its effectiveness in accurately identifying and classifying different emotional states based on facial features. The robust architecture of the FER model and the learning process allow it to distinguish the nuances of facial expressions, contributing to a more detailed understanding of human emotions. Consequently, such a system could be useful for medicine, social research, marketing, and entertainment, helping to analyze emotional responses and improve user interaction strategies.

Keywords: neural network; convolution; JavaScript; TensorFlow.

Issue
No 6 (2023)
Section
Articles