How does it all work? In this blog, I explain the process behind my Sudoku Grab app, a solution that uses basic image processing techniques to recognize Sudoku puzzles. This is done by locating the puzzle in an image, turning it back into a square form, segmenting it to find potential numbers, and lastly, recognizing those numbers. This involves simple thresholding techniques, blob extraction algorithm, perspective transform, and a Neural Network for Optical Character Recognition (OCR) to recognize digits from the photograph. I conclude by mentioning the potential for multiple enhancements to this process. 14 August 2009