b13_enhanced_OCR

Description

This project is about enhancing Optical Character Recognition (OCR) using machine learning. The main script is website_recognition.ipynb which uses a trained model to predict digits from photos taken and drawings both from the website specially built for this project.

Usage

Create a conda environnement

Once you retrieved this porject, execute the following command:

conda create --name <env> --file requirements.txt

where <env> is the name you choose for this new environnement

Train model

You first need to locally run notebooks/number_recognition.ipynb which will train, validate and test a Sequential model with MNIST dataset. This model will then be saved as a keras model in models/number_recon_model.keras

jupyterlab notebooks/number_recognition.ipynb

Take your own photos

Thanks to Loke-60000's and fdeage's website including in this project, you can take photos of handwritted numbers. To do so :

cd website
python -m http.server 8001

Your photos and drawings will be saved in your downloads' folder. You need to copy them into this_project/data/webcam for the photos and this_project/data/drawings for the drawings. Each file must be renamed with its own value (if several files have the same value, add underscores ("_") after the value). This will serve to display the actual value of the number and calculate the accuracy of the prediction.

Make predictions

Run website_recognition.ipynb to load your images, process them and use number_recon_model.keras to predict what number you draw.

jupyterlab website_recognition.ipynb