/sudoku_DL

An implemetation of the Sudoku Solver using Deep Learning

Sodoku Solver using Deep Learning

The learning based method given in this blog and the non-learning based method that I tried here were not very accurate. This project uses the LeNet Convolutional Neural Network for digit recognition to extract the digits from the grid of the sudoku puzzle.

How to Setup

This project uses the Caffe framework to run the CNN.

  1. Download the latest Caffe version here.
  2. Copy the contents of models/ folder into the caffe/models/ folder.
  3. Copy the contents of examples/ folder into the caffe/examples/ folder.
  4. Build Caffe according to the instructions given on their website.
  5. From the caffe/ folder, run the following command
./build/examples/sudoku/sudoku.bin

Running for different images

The image file is set in caffe/examples/sudoku/sudoku.cpp in line number 570.

Edit the line, build caffe again and run the command.

Implementation

You can read more about the implementation in my blog here.

TODO

  • Improve the accuracy by finetuning with more data.