Information system of semantic search of the image to the text of advertising announcement

Abstract

In connection with the rapid development of information technology, modern society is in need of solving the problems of the accumulation, storage and automated processing of semiotically heterogeneous arrays of texts. An important task of the developed technologies is the search for operational methods for issuing a wide variety of information in accordance with the user’s request. The aim of the proposed work is to develop a computer information system that allows you to select the most appropriate illustration for the given text of the advertisement. In this case, the main content of the text is proposed to be presented in the form of a certain set of main and secondary keywords, and the content of the illustration in the form of a multi-level complex of descriptors. The proximity of these contents (and, accordingly, the correlation between the image and the verbal component of the advertisement) can be determined through the maximum number of matches of support words and descriptors, taking into account their importance. An important feature of this system is the use of a thesaurus method of representing the image content, which is a description of the illustration in the form of a certain set of words and the relations between these words within a certain subject area.

Keywords

• information system
• semantic search
• advertisement
• computer model
• keyword
• main content
• thesaurus description