HUMAN EMOTION RECOGNITION SYSTEM USING DEEP LEARNING ALGORITHMS
DOI:
https://doi.org/10.30837/ITSSI.2022.21.060Keywords:
object detection, classification of objects, supervised learning, recognition of emotionsAbstract
The subject of research in the article is the software implementation of a neural image classifier. The work examines emotions as a special type of mental processes that express a person’s experience of his attitude to the surrounding world and himself. They can be expressed in different ways: facial expressions, posture, motor reactions, voice. However, the human face has the greatest expressiveness. Technologies for recognizing companies to improve customer service use human emotions make decisions about interviewing candidates and optimize the emotional impact of advertising. Therefore, the purpose of the work is to find and optimize the most satisfactory in terms of accuracy algorithm for classifying human emotions based on facial images. The following tasks are solved: review and analysis of the current state of the problem of "recognition of emotions"; consideration of classification methods; choosing the best method for the given task; development of a software implementation for the classification of emotions; conducting an analysis of the work of the classifier, formulating conclusions about the work performed, based on the received data. An image classification method based on a densely connected convolutional neural network is also used. Results: the results of this work showed that the method of image classification, based on a densely connected convolutional neural network, is well suited for solving the problems of emotion recognition, because it has a fairly high accuracy. The quality of the classifier was evaluated according to the following metrics: accuracy; confusion matrix; precision, recall, f1-score; ROC curve and AUC values. The accuracy value is relatively high – 63%, provided that the data set has unbalanced classes. AUC is also high at 89%. Conclusions. It can be concluded that the obtained model with weights has high indicators of recognition of human emotions, and can be successfully used for its purpose in the future.
References
Hess, U., Thibault, P. (2009), "Darwin and Emotion Expression", American Psychologist, Vol. 64, No. 2, P. 120–129. DOI: https://doi.org/10.1037/a0013386
Turabzadeh, S. (2018), "Facial expression emotion detection for real-time embedded systems", Technologies, No. 6 (1), 17 p. DOI: https://doi.org/10.3390/technologies6010017
Ekman, P. (1970), "Universal facial expressions of emotion", California Mental Health Research Digest, Vol. 8, No. 4,
P. 151–158.
Viola, P., Jones, M. (2001), "Rapid object detection using a boosted cascade of simple features", Proceedings of
the 2001 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. CVPR 2001, Kauai, HI, USA,
–14 December, P. 1–9.
Coots, T., Edwards, G., Taylor, Ch. (2001), "Active Appearance Model", IEEE Transactions on Pattern Analysis and
Machine Intelligence, Vol. 23, No. 6, P. 681–685.
Coots, T., Taylor, K. (2004), "Statistical appearance models for computer vision", Technical Report, University of Manchester, Wolfson Image Analysis Group, Imaging Science and Biomedical Engineering, Manchester M13 9PT, United Kingdom, 125 p.
Van Dyke, Parunak H. (1995), Neural Networks for Pattern Recognition. New York: Christopher M. Bishop, Oxford University Press, 482 p.
Yashina, E., Artiukh, R., Pan, N., Zelensky, A. (2019), "Information technology for recognition of road signs using
a neural network", Innovative technologies and scientific solutions for industries, No. 2 (8). DOI: https://doi.org/10.30837/2522-9818.2019.8.130
Mehryar, M., Rostamizadeh, A., Talwalkar, A. (2018), Foundations of Machine Learning, Cambridge: The MIT Press, 505 p.
Krizhevsky, A., Sutskever, I., Hinton, G. E. (2017), "Image Netclassification with deep convolutional neural networks", Communications of the ACM, Vol. 60, No. 6, P. 84–90.
Huang, G., Sun, Y., Liu, Z. (2016), "Deep networks with stochastic depth", European Conf. on Computer Vision (ECCV), Amsterdam, Netherlands, October, P. 646–661.
Srivastava, R. K., Greff, K., Schmidhuber, J. (2015), "Training very deepnetworks", Advances in Neural Information Processing Systems 28 (NIPS 2015), Montreal, Canada, December, P. 2377–2385.
He, K., Zhang, X., Ren, S. (2016), "Deep residual learning for image recognition", The IEEE Conf. on Computer Vision
and Pattern Recognition (CVPR), Las Vegas, NV, USA, 26 June – 1 July, P. 1–9.
Larsson, G., Maire, M., Shakhnarovich, G. (2017), "Fractalnet: ultra-deepneural networks without residuals", 5th International Conference on Learning Representa-tions, Toulon, France, April, P. 1–11.
Gao, H., Zhuang, L., Laurens van der Maaten, Kilian, W. (2017), "Weinberger Densely Connected Convolutional Networks", The IEEE on Conference on Computer Vision and Pattern Recognition (CVPR), Honolulu, HI, USA, 21–26 July, P. 1–9.
FER2013 (Facial Expression Recognition 2013 Dataset), available at: https://paperswithcode.com/dataset/fer2013 (last accessed 12.05.2022).
Downloads
Published
How to Cite
Issue
Section
License
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
Our journal abides by the Creative Commons copyright rights and permissions for open access journals.
Authors who publish with this journal agree to the following terms:
Authors hold the copyright without restrictions and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License (CC BY-NC-SA 4.0) that allows others to share the work with an acknowledgment of the work's authorship and initial publication in this journal.
Authors are able to enter into separate, additional contractual arrangements for the non-commercial and non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgment of its initial publication in this journal.
Authors are permitted and encouraged to post their published work online (e.g., in institutional repositories or on their website) as it can lead to productive exchanges, as well as earlier and greater citation of published work.
