Chess piece location and identification using Machine Learning

Abstract

This project aims to identify and locate chess pieces from a chess board photograph to recreate the chess board position in digital format. To do so, several approaches will be tested to challenge previous limitations and requirements for this task to overcome them and achieve a simpler and more flexible solution. Overcoming said limitations are of great importance, as doing so will increase accessibility to chess recognition models, moving one step closer to end-user making automatic digital recreations of their over-the-board chess games.