The image data and student database used in this project is closed source and is not provided in this repository. To get access to sample data you need to make a request here.
You can also use the skeletan image provide in data directory to create your own sample images, and fill up the CSV file with corresponding data along with random names and roll numbers for testing. You need to move your sample images in data>images.
- First you need to ensure following directory structure in your google drive
├─ My Drive
└─ SAFE_OCR
└─ data
└─ models
- (optional) Run MNIST_model_training.ipynb and EMNIST_Model.ipynb in google colab. This should save trained model in models>MNIST_Model directory and models>EMNIST_Model respectively. However, trained models are already provided in the repository.
- Run SAFE_OCR.ipynb in google colab.