As theano
has been discontinued and TensorFlow
has taken over. I am moving this project to
TensorFlow. So this work is discontinued. Code duplication from the rnn_ctc
library for
Reccurent Nueral Networks
with Connectionist Temporal Classification
has been deleted.
The code for 'scribe'ing Indian Language Text has also been moved to a new package
IndicScribe
The Chamanti OCR based on TensorFlow
and IndicScribe
will be up in my repositories.
This project aims to build a very ambitious OCR framework, that should work on any language. It will not rely on segmentation algorithms (at the glyph level), making it ideal for highly agglutinative scripts like Arabic, Devanagari etc. We will be starting with Telugu however. The core technology behind this is going to be Recurrent Neural Networks using CTC from the repo rnn_ctc.
- numpy
- scipy
- theano
You can read the developer documentation for more details about the code and configurations.