An OCR project to recognize texts present in an image. Individual english characters and numbers are recognized using a neural network.
(Currently the OCR works on only document images with font in black color and everything in white. If proper pre-processing is done then it will work for other types of image as well.)
The neural network is trained using 62K examples. The examples consist of letter a-z, A-z and digits 0-9. (Code of neural network is present in neural_network folder)