Identification of human emotions using the Keras and TensorFlow neural Network.

  • V. Melnyk LNTU
  • E. Melnyk LNTU
  • B. Shulga LNTU
Keywords: modeling, identification, human emotions, convolutional neural network, KERAS


This article presents the results of research on the determination of human emotions through neural networks. The development of the model for image analysis was carried out using TensorFlow, and the training was implemented using Keras. The input was used from the - FER2013 archive. The OpenCV library was used to analyze the images. Programming language - Python 3. This toolkit is considered the most popular and most convenient for building neural networks, as well as systems of deep learning. Neural networks and machine learning are the most popular technologies at the moment. Particularly great results can be achieved by combining this technology with other known - for example, technology object-oriented programming. This combination of technologies has a wide range of applications in various fields, ranging from ordinary photo luminaries outlined in social networks, and ending with the control of the behavior of citizens of the state or even the planet. The analysis of emotions allows grocery and advertising companies to significantly increase sales, which in turn will increase profits [12]. Being a brilliant interlocutor, having the ability to manipulate people knowing what they think, conduct stress tests of employees and evaluate their reactions, determine the human reaction to advertising, announcement, speech to collect, process and draw conclusions


Neural Networks and Genetic Algorithms - K .: Korniychuk,. 2008. - 446 p.

Deep learning in Python - JV: Peter,. 2018 - 400 p.

Neural Networks: Full Course - M: Dialectics-Williams,. 2017 - 1104 p.

TensorFlow Official Documentation. URL: (accessed: 03/18/2018)

Keras Official Documentation. URL: (Accessed: 03/18/2018)

Platform for machine learning and predictive modeling competitions. URL: (accessed: 03/18/2018)

OpenCV Official Documentation. URL: (Accessed: 03/18/2018)

Artificial Intelligence Programming in Applications - DMK Press, M .: Tim Jones,. 2006. - 312 c.

Neural Networks: Recognition, Management, Decision Making - M: Finance and Statistics,. 2004. - 176 p.

Basic concepts of neural networks - M.: Williams,. 2003. - 288 p.

Fundamentals of Data Science and Big Data. Python and data science - JV: Peter, 2016. - 336 p.

Python for complex tasks: data science and machine learning - JV. : "Peter",. 2016 - 576 p.

Wiener N. Cybernetics, or Control and Communication in Animal and Machine. / Trans. with English. IV Solovyov and G.N. Povarova; Ed. GN The cook. - 2nd edition. - M .: Science; Main edition of publications for foreign countries, 1983. - 344 p.

Galeshchuk S. Artificial neural networks in forex market forecasting.

Sedai AV Use of neural networks for modeling and forecasting of financial activity of transport enterprise.

Traffic from January 2014 to September 2017, by month [Electronic resource]

Vikas P. Deshpand. An Evaluation of Naive Bayesian Anti-Spam Filtering Techniques / Vikas P. Deshpande, Robert F. Erbacher, Chris Harris // Proceedings of the 2007 IEEE Workshop on Information Assurance United States Military Academy, West Point, 2007 - NY June 20-22 .

Graham P. A Plan for Spam / P. Graham, 2002. [Electronic resource] - Access mode: (accessed 10/06/2019).

Melnyk KV, Melnyk VM, Koptyuk Yu.Yu. Investigation of neural network image recognition methods. Computer-Integrated Technology Science: Education, Science, Production, Lutsk, 2019. № 35. - p. 161-165.

Goncharov DA Acoustic Information Recognition Software in Sensor Networks / Master's Thesis for Master's Degree in Specialty 123 "Computer Engineering" in Specialized Computer Systems. Kiev, KPI. 2018 - 122 p. URL: bitstream / 123456789/25805/1 / Matviiv_magistr.pdf.

Sagan V.Yu. Biomedical signal analysis system for driver status monitoring. / Master's Thesis for Master's Degree in Specialty 122 Computer Science and Information Technology. // Kiev, KPI. 2018 - 122 p.

A.A. Yarovyi, SG Kashubin, OO Sandpiper. Recognition of facial expressions of human face. / Technical vision and artificial intelligence systems with image processing and stratification. Vinnitsa. - VNTU, 2015. - p.76-83.

Abstract views: 25
PDF Downloads: 9
How to Cite
Melnyk, V., Melnyk, E., & Shulga, B. (2019). Identification of human emotions using the Keras and TensorFlow neural Network. COMPUTER-INTEGRATED TECHNOLOGIES: EDUCATION, SCIENCE, PRODUCTION, (36), 109-122.
Computer science and computer engineering