A computer vision application that identifies pieces on a chess board from real-life photos.
This is a computer vision application that identifies pieces on a chess board from real-life photos and converts it to FEN (Forsyth–Edwards Notation). It also generates a link to the analysis board on Lichess, where the pieces are pre-set according to the user-uploaded image. You can use it you quickly save real-life chess games for later continuation, as well as to get move suggestions from different chess engines such as Stockfish.
The animated gif above illustrates the functionality of the application, showcasing its ability to identify pieces on a chessboard from real-life photos and convert them into FEN code. The link generated leads to the Lichess analysis board with the pieces already arranged according to the user-uploaded image.
-
Download the chess-pieces detection YOLO model from from this link. (the file is too big to host on github)
-
Open app.ipynb and enjoy the code.