Analysis of Meaning Attribution During the Embedding Process in Computer Vision and Natural Language Models Applied to the Urban Environment

• Grant number: 25/00043-9
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: April 01, 2025
• End date: March 31, 2026
• Field of knowledge: Physical Sciences and Mathematics - Computer Science
• Principal Investigator: Roberto Marcondes Cesar Junior
• Grantee: Mateus Matzkin Gurza
• Host Institution: Instituto de Matemática e Estatística (IME). Universidade de São Paulo (USP). São Paulo , SP, Brazil
• Associated research grant: 22/15304-4 - Learning context rich representations for computer vision, AP.TEM

Abstract

This scientific initiation project investigates the use of computer vision and artificial intelligence to analyze urban scenarios, focusing on the detection and description of entities in complex environments. The objective is to evaluate how different models represent the semantics of images and, also, to identify the information loss during the encoding process into embedding vectors and the generation of textual descriptions. Using real data from urban environments and their specific characteristics, the research aims to contribute to the development of smarter, more accessible, safer, and sustainable cities, optimizing the use of AI in urban planning and management.