To develop a premise for the task that this repository intends to preform, read TextRecognition-CRNN.pdf

This repository consists of two directories - TextImageGenerator and CRNN. The TextImageGenerator is used to generate text images from a dictionary that you may provide. On the other hand, CRNN contains the code-base for training and testing a CRNN model to recognise/transcribe the text in images.

To get started, clone this repository and install the dependencies listed in requirements.txt. Alternatively, just run pip install -r requirements.txt to install all the required dependencies.

To generate the data, follow steps mentioned in the readme present in TextImageGenerator.

To start training your own model OR for testing a trained model, follow steps mentioned in the readme present in CRNN