Book cover
International Conference on Intelligent Information Technologies for Industry

IITI 2022: Proceedings of the Sixth International Scientific Conference “Intelligent Information Technologies for Industry” (IITI’22) pp 142–151Cite as

Sift Descriptor for Social Media User Accounts Matching
Anastasia A. Korepanova & Maxim V. Abramov
Conference paper
First Online: 31 October 2022
12 Accesses

Part of the Lecture Notes in Networks and Systems book series (LNNS,volume 566)

Abstract
The task of user accounts matching in different social media and determining those belonging to the same user is relevant in various contexts related to the analysis of social media. The solution of this problem is of both theoretical importance and allows you to expand the understanding of the behavior of users in social media, as well as practical and can be applied to collect data about a single user. This work is devoted to solving the problem of automation of matching the accounts of social media users by analyzing the graphic content posted in the accounts. Previously, a method was proposed to solve this problem, based on the use of a number of content features, the current study develops the proposed approach by searching for duplicate images using SIRF. As a result, a new model for classifying pairs of accounts into the classes “belongs to one user” and “does not belong to one user” was proposed, which achieved higher accuracy and f-score compared to previous results.