AI Chess Project
Making Chess Accessible for the Visually Impaired
About the Project
My software uses neural networks coded in Python that have been trained to recognize a chess board and converted the position of chess pieces into chess notation that could be understood by any chess player. I approached several professional visually impaired chess players to understand the issue they faced and how I could tackle it best. Through these conversations, I learned that the entire process had to be connected to a keybind and had to be running on a background application. After a year of work, I created an application with a basic GUI framework on TKInter that ran in the background. After the click of a shortcut (Alt+3), the application took a screenshot of the user’s screen, ran the image through my program on a cloud server I had set up so the process was quick and efficient on all kinds of laptops, and immediately outputted the chess notation onto a notepad file that popped up on the user’s screen, which the visually impaired user could read through their screen reader. After testing this software with professional visually impaired chess players, they were extremely thankful and pleasantly surprised at how well the program worked with all kinds of chess resources, whether it be applications like ChessBase, websites like Chess.com, or blogs and YouTube videos with chess boards.
As Featured on ChessBase India:
Instructions to Install and Run the program
Extract the zip file
Install the application
Make sure screenreaders such as JAWS are already running in the background
The project should now start running in the background/inside your taskbar
Make sure the chess board is covering over 50% of the screen
Make sure the chess board colours are differentiable, any standard colours like white and brown, white and black, white and green should work
When you want to get the notation from the chess position on the screen, run the keybind 'Alt + 3'
Within a few seconds a notepad file should open with the chess notation