Recognition of formation patterns in sports visual data

Author nameKonstantinos Stavrianos
Title
Recognition of formation patterns in sports visual data
Year2021-2022
Supervisor

Ilias Maglogiannis

IliasMaglogiannis

Summary

Sport analytics is a broad field in artificial intelligence and is a part of our daily life. Each sport has its own rules and philosophy so decisions vary depending on the occasion. Also the AI algorithms cannot be applied to every scenario I come from the real world of sports because the decisions have to be specific. We should be able to vectorize the problem in order to arrive at conclusions that have decision value. In the analysis of football visual data a need has been identified to implement an algorithm for use by football experts, analysts and coaches. The idea of creating an artificial intelligence model that recognizes through machine learning decisions if the formation is the particular one we have chosen. In our scenario the models first identify the footballers of our own team and then make a decision whether the formation is 4-4-2 or not.