Analysis of Visual Information to Determine the Emotional State of People

The demand for emotion recognition software in various fields of human endeavour is undeniable. Research has focused on the application of various technologies and methods, including machine learning, neural networks and big data analytics. Available software in the field of emotion recognition is characterised by a number of drawbacks: high cost; not always transparent data privacy and ethics; dependence on data quality; and narrow specialisation of data processing. Thus, it is relevant to develop the software taking into account the minimisation of the above disadvantages. In implementing the application, data collection and preparation was carried out, a recognition model was developed using neural networks and machine learning, and the accuracy of the model was evaluated. The application implements user authorisation using encryption algorithms, file uploading, file analysis, saving and viewing the results. The developed software allows recognising six basic emotions: joy, sadness, anger, fear, surprise and neutral state.