Search academic texts

Information × Registration Number 0223U002319, 0121U110662 , R & D reports Title Methods, models and computer tools for detecting destructive influences in the media space popup.stage_title Head Tereikovskyi Ihor A., Доктор технічних наук Registration Date 15-02-2023 Organization National Technical University of Ukraine "Igor Sikorsky Kyiv Polytechnic Institute" popup.description2 The object of research is the processes of recognizing destructive influence in multimedia messages of the mass media. The purpose of the work is to provide effective recognition of destructive influence in multimedia messages of mass media. Research methods – methods of the theory of digital signal processing, neural networks, simulation modeling and expert evaluation were used. It is proposed to implement the assessment of the destructive impact of viewing multimedia messages of the mass media through the use of neural network means of recognizing the emotional state of a person from the image of a face. The imperfection of such recognition tools is shown from the point of view of their insufficient adaptation to the conditions of use. As a result of the conducted research, a method of semantic segmentation of images using a convolutional neural network was developed, which involves the adaptation of the encoder and decoder to significant application conditions. The obtained experimental results show that the application of the proposed method allows, avoiding complex long-term experiments, to build a neural network model that, with a sufficiently short training period, ensures the achievement of facial image segmentation accuracy of about 0.8, which corresponds to the best systems of a similar purpose. Product Description popup.authors Tereikovska Liudmyla Tereikovskyi Oleh popup.nrat_date 2023-02-15 Close