Rooba
/
Kindai-OCR
mirror of https://github.com/ducanh841988/Kindai-OCR.git
1
1
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Update README.md

This commit is contained in:
LE DUC ANH 2020-07-08 11:41:37 +09:00 committed by GitHub
parent 2cafcde9a8
commit 83662fd9a1
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 4 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
README.md
View File

@ -6,9 +6,12 @@ OCR system for recognizing modern Japanese magazines
This repo contains an OCR sytem for converting modern Japanese images to text. This repo contains an OCR sytem for converting modern Japanese images to text.
This is a result of [N2I project](http://codh.rois.ac.jp/collaboration/#n2i) for digitization of modern Japanese documents. This is a result of [N2I project](http://codh.rois.ac.jp/collaboration/#n2i) for digitization of modern Japanese documents.
The system has 2 main modules: text line extraction and text line recognition. The overall architechture is shown in the below figure. The system has 2 main modules: text line extraction and text line recognition. The overall architechture is shown in the below figures.
![alt text](https://github.com/ducanh841988/Kindai-OCR/blob/master/images/TextlineExtraction.jpg "text line extraction")
For text line extraction, we retrain the CRAFT (Character Region Awareness for Text Detection) on 1000 annotated images provided by Center for Research and Development of Higher Education, The University of Tokyo. For text line extraction, we retrain the CRAFT (Character Region Awareness for Text Detection) on 1000 annotated images provided by Center for Research and Development of Higher Education, The University of Tokyo.
![alt text](https://github.com/ducanh841988/Kindai-OCR/blob/master/images/TextlineRecognition.jpg "text line recognition")
For text line recognition, we employ the attention-based encoder-decoder on our previous publication. We train the text line recognition on 1000 annotated images and 1600 unannotated images provided by Center for Research and Development of Higher Education and National Institute for Japanese Language and Linguistics, respectively. For text line recognition, we employ the attention-based encoder-decoder on our previous publication. We train the text line recognition on 1000 annotated images and 1600 unannotated images provided by Center for Research and Development of Higher Education and National Institute for Japanese Language and Linguistics, respectively.