ocropus-archive / DUP-ocropy

Python-based tools for document analysis and OCR
Apache License 2.0
3.42k stars 590 forks source link

Could you recommend some materials about the algorithm you use? #31

Closed hsmyy closed 7 years ago

hsmyy commented 9 years ago

Hi! I feel this project is very interesting and I want to learn from it. So could you recommend me some materials(papers or books) you referred in this project? Thank you very much

adnanulhasan commented 9 years ago

There is one paper "High Performance OCR for Printed Latin and Fraktur using LSTM Networks". Please see this paper and references therein to learn more about the recognition engine used in ocropy.

hsmyy commented 9 years ago

Thank you very much for your help!

nixar commented 9 years ago

It's just a quick google away, but to save everyone 10s, here's a link to the preprint: http://www.cis.uni-muenchen.de/~springmann/papers/2014-04-07-DATeCH2014-Springmann.pdf

zuphilip commented 9 years ago

That is a different paper with different authors. Google shows you here a paper refering to the one mentioned above. The official published version can be found on IEEE Xplore: http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=6628705

nixar commented 9 years ago

Ah yes you're right. I had saved it for later reading and not paid attention. Anyone has a link to a non-paywalled version?

adnanulhasan commented 9 years ago

Here is the link of my website to see the pdf version: https://docs.google.com/a/iupr.com/file/d/0B_81tRo0ZXnLSE9BeGdNcDBXYms/edit

It's a slightly different version, but essentially it is the same paper .

adnanulhasan commented 9 years ago

There were multiple requests to share the paper. Here is the link of this paper on Academia: https://www.academia.edu/3568100/High_Performance_OCR_for_English_and_Fraktur_using_LSTM_Networks I hope that viewing this paper here will be easier.

adnanulhasan commented 9 years ago

This issue is resolved, please close it if you are satisfied. Thanks!

tianyaqu commented 8 years ago

I would like to recommend this paper "A Novel Connectionist System for Unconstrained Handwriting Recognition",it explains in every details about the LSTM model and CTC used in this project. Here is link of this paper: www.cs.toronto.edu/~graves/tpami_2009.pdf Hope it may help.

zuphilip commented 7 years ago

We are collecting some publications around ocropus on the wiki page https://github.com/tmbdev/ocropy/wiki/Publications . Everybody can add more publications there, maybe we should also add a more general section about the networks used in the software.