Análises de competições presentes na Plataforma Kaggle para auxiliar no desenvolvimento de novas soluções para problemas de visão computacional

Abstract

This work aims to analyze and discuss previously collected data regarding computer vision problems and their best solutions on the Kaggle platform. The purpose of these discussions was to try to understand what are the most used approaches in some types of problems and why they are used more often than others. To accomplish that, before the development of this monography, a research project was carried out to collect data about the main characteristics of some computer vision competitions on Kaggle’s platform. For this work, relevant questions were raised about the development of new solutions to computer vision problems. Each question was widely discussed and substantiated with an analysis of the data collected in the research project and concepts found in the literature. Finally, it concludes the importance of the analysis of existing solutions to assist in the discussion of certain steps of the process of development of solutions to new problems.