Free Indic OCR

From WorkOutWiki2009

Contents 1 Proposer 2 Purpose 3 Pre-requisites 3.1 Getting and compiling the code 3.2 Links to overall design/architecture 3.3 Tasks 3.4 Existing work 3.5 Getting in touch 4 Participants

Proposer

Debayan Banerjee

debayanin <att> gmail <dott> com

Purpose

After some development efforts, we have an OCR engine that works well http://code.google.com/p/tesseractindic. The aim now is to add support for as many languages as possible. Also, a GUI and Web frontend to the OCR system is desirable so that end-users can use it without having to deal with the command line.

Pre-requisites

Knowledge of any of the below 3 will do.

 Python / PyGTK
 C++
 Django

Going through http://hacking-tesseract.blogspot.com/ will help.

Getting and compiling the code

Read the instructions at http://code.google.com/p/tesseractindic.

Links to overall design/architecture

 http://code.google.com/p/tesseract-ocr/w/list
 http://tesseract-ocr.googlecode.com/files/TesseractOSCON.pdf

Tasks

 1> Add training data for new languages. So far, only Hindi and Bengali are supported. Read http://code.google.com/p/tesseractindic/wiki/TrainerGUI.
 2> Integrate TesseractIndic with OCRFeeder http://live.gnome.org/OCRFeeder.
 3> Create a web framework for the OCR using Django.

Existing work

 http://code.google.com/p/tesseractindic/downloads/list

 http://crblpocr.blogspot.com/

Getting in touch

 Indlinux mailing list

 #indlinux and #sarai on irc.freenode.net

Participants

 # Santhosh Thottingal

 # Sayamindu Dasgupta