Development and Application of Artificial Intelligence Techniques for Image Analysis

• Grant number: 25/12082-9
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: July 01, 2025
• End date: June 30, 2026
• Field of knowledge: Interdisciplinary Subjects
• Principal Investigator: Sylvia Iasulaitis
• Grantee: Heitor Gomes de Oliveira
• Host Institution: Centro de Educação e Ciências Humanas (CECH). Universidade Federal de São Carlos (UFSCAR). São Carlos , SP, Brazil
• Associated research grant: 22/03090-0 - Analysis of large amounts of political data and complex networks: mining modelling and applications in Computational Political Science, AP.PNGP.PI

Abstract

With the growing use of social media platforms, various types of content are shared daily, especially memes. Memes are a type of composite content, generally consisting of text and an image, characterized as a form of rapid communication to express ideas, criticism, humor, and more. However, this type of artifact has also been used to convey messages with harmful content. Due to their multimodal nature (text and image), the automatic analysis of memes is a challenging problem, as understanding their content is only possible when the text and image are analyzed jointly. This also requires external knowledge related to the semantic and cultural context in which the meme is embedded. Therefore, one of the crucial steps is the binary classification of an image as a meme or not. In this sense, the present research proposal aims to build/adapt an image classification model. This model will assist in creating a database of memes in Brazilian Portuguese to be used in research conducted by the study group "Interfaces," related to hate speech, sentiment analysis, harmful content detection, among other related studies. (AU)