Skip to content

Latest commit

 

History

History
14 lines (7 loc) · 572 Bytes

README.md

File metadata and controls

14 lines (7 loc) · 572 Bytes
Raw

Optical Character Recognition

Optical Character Recognition (OCR) implemented with Convolutional Neural Network (CNN) in TensorFlow.

Usage

Download the dataset in the link below to this folder and execute the "ocr_deep.py" file.

Data set

The Chars74K dataset has been used to train this model. To train your model, it can be downloaded from http://www.ee.surrey.ac.uk/CVSSP/demos/chars74k/.

For this implementation the computer fonts version has been used, but it may work with the other versions too, you just need to change the "data_providers.py" path.