Chess board state detection and move recommendation system
Our capstone project aims to create an end-to-end comprehensive chess analysis tool that integrates digital insights into over the board play in chess. The project's primary objective is to utilize Computer Vision (CV) technology to interpret 3-D images of chess boards, converting these into 2-D board states using Forsyth–Edwards Notation (FEN).
Subsequently, we will process the position and provide the user with chess insights to better help them navigate the position as they are playing or analyzing. We plan to build two types of chess engine systems.
The primary focus of the first system will be matching scanned board states with similar positions from a database of master games. The second system will be reinforcement learning based using AlphaZero algorithm. The project intends to bridge the gap between offline chess and online analysis resources, providing users with gameplan inspiration through recognition of chess positions and advanced game analysis, offering a powerful resource for both enthusiasts and serious chess players.
Chess Games
Computer Vision