Application of Convolutional Neural Networks for the Classification of Human Emotions




Abstract:
Computer vision training is implemented using deep learning. Deep learning can be applied to solve a variety of classification problems. Accordingly, the problem of recognizing human emotions based on facial images can be singled out. For this problem, the use of convolutional neural networks proved to be the best solution. Specifically, the convolutional neural network models used in this paper are ResNet50, VGG16, and VGG19. The CNN model was implemented using the Keras library and the Python programming language. As input to these models, the RAF-DB database containing images of human faces with emotions was used. Based on the results obtained from the mentioned three CNN models, VGG16 proved to be the best, achieving a precision of 83.61%. VGG19 was on the second place with an accuracy of 82.37%, and the worst was ResNet50, whose accuracy was 75.31%.

CITATION:

IEEE format

N. Matijašević, A. Samčović, M. Đogatović, “Application of Convolutional Neural Networks for the Classification of Human Emotions,” in Sinteza 2023 - International Scientific Conference on Information Technology, Computer Science, and Data Science, Belgrade, Singidunum University, Serbia, 2023, pp. 52-59. doi:10.15308/Sinteza-2023-52-59

APA format

Matijašević, N., Samčović, A., Đogatović, M. (2023). Application of Convolutional Neural Networks for the Classification of Human Emotions. Paper presented at Sinteza 2023 - International Scientific Conference on Information Technology, Computer Science, and Data Science. doi:10.15308/Sinteza-2023-52-59

BibTeX format
Download

RefWorks Tagged format
Download