https://gitlab.com/space-kerala/lekha-OCR
Note: This project is under development. All modules are written in python. For image processing opencv is used
Printed text recognizer for Malayalam, Lekha OCR is an optical character recognizer trained for the recognition of printed malayalam Documents.
OpenCV 2.4.11 Python 2.7.9
Currently only block recognition is available. Layout analysis is at devoloping stage. To recognize a scanned malayalam document and get the malayalam characters as output.
$./lekhaocr <filename>
for example
$./lekhaocr Example/dc_books_page.png
This project is funded by ICFOSS, technically guided under space-kerala.
Arun Joseph contributed most of the engine devolopments. Jithin Thankachan contributed some additional features, training tool and helped in documentatiuon. Rijoy V contributed in initial research. Ambily Sreekumar contibuted in building data set for training. Arun M helped in project mangement and technical assistace.