OpenCV- OCR project

1. A pdf needs to be provided as an input.

2. It's broken down to images(page-wise).

3. From those images, text is detected.

4. Text detection is done by creating bounding boxes.

5. Coordinates of these BBoxes are stored in a JSON file.

In the notebook, it states 'hin.traineddata' is used however, I had changed it to 'san.traineddata'.

For better accuracy, one can make use of kernels too.