This is a simple fun play python GUI app to solve the Soduku using the input in the form of image , Video or the Webcam ( both from the system webcam or by using the ip webcam ) . It is build on the top of Keras API . The trained model ( ocr/model.hdf5 ) takes the picture of each digit and predict it accordingly .
Opencv is highly used in it to find the block Inside the picture or the video
What could have been a perfect way to utilise the lockdown period ? Like most of my time in painting , Netflix . I thought of to start with the Deep Learning . After completing the Artificial Neural Network . I move on to CNN .So this is under that I made .
The Project is divided into three parts:
1-> Preprocessing the data, Finding the outer box of sudoku , Using waped transform ( like that in scanner ) , Forming a hough Lines .
2 -> Forming a number image . i. e . it containg only the numbers , Then forming the 2 - Dimension array by predicting each digits inside .
3 -> Solving the sudoku and , Then finally using insceptive wraping to print the answer on the original frame.
The code is written in python 3.7 . It you don't have installed you can find it on google . If you have a lower version of Python you can upgrade using the pip package , ensuring you have the latest version of pip . To install the required Packages and libraries , run teh command in the project directory after cloning the repository.
pip install -r requirements.txt
After following the above steps of installation . Open the terminal( cmd, powershell ) in the project directory and use the command
python Gui_Final_Sudoku_solver.py
Numpy ( for n-dimension array )
PIL ( for image manipulation )
Keras ( To train the model )
Tensorflow-gpu ( For deep learning )
Tensorboard ( Visual analysis os model )
Opencv-python ( Scientific library for Image Related stuffs )
Tkinter ( GUI for the program )