Rooba
/
Kindai-OCR
mirror of https://github.com/ducanh841988/Kindai-OCR.git
1
1
Fork 0
Code Issues Packages Projects Releases Wiki Activity
OCR system for recognizing modern Japanese magazines
24 Commits 4 Branches 0 Tags 132 MiB
Python 100%
Go to file
LE DUC ANH 89b04e0f1b
Update README.md 		2020-08-05 15:42:03 +09:00
basenet initial commit 2020-07-08 11:31:26 +09:00
data add code for both GPU and CPU 2020-07-22 11:01:12 +09:00
images add images 2020-07-08 11:34:55 +09:00
pretrain initial commit 2020-07-08 11:31:26 +09:00
.gitignore initial commit 2020-07-08 11:31:26 +09:00
README.md Update README.md 2020-08-05 15:42:03 +09:00
coordinates.py initial commit 2020-07-08 11:31:26 +09:00
craft.py initial commit 2020-07-08 11:31:26 +09:00
craft_utils.py initial commit 2020-07-08 11:31:26 +09:00
data_loader.py initial commit 2020-07-08 11:31:26 +09:00
decoder.py add code for both GPU and CPU 2020-07-22 11:01:12 +09:00
encoder.py initial commit 2020-07-08 11:31:26 +09:00
encoder_decoder.py add code for both GPU and CPU 2020-07-22 11:01:12 +09:00
evaluation.py initial commit 2020-07-08 11:31:26 +09:00
file_utils.py add code for both GPU and CPU 2020-07-22 11:01:12 +09:00
gaussian.py initial commit 2020-07-08 11:31:26 +09:00
imgproc.py initial commit 2020-07-08 11:31:26 +09:00
mep.py initial commit 2020-07-08 11:31:26 +09:00
mseloss.py initial commit 2020-07-08 11:31:26 +09:00
requirements.txt add code for both GPU and CPU 2020-07-22 11:01:12 +09:00
test.py add code for both GPU and CPU 2020-07-22 11:01:12 +09:00
torchutil.py initial commit 2020-07-08 11:31:26 +09:00
translate_line.py initial commit 2020-07-08 11:31:26 +09:00
utils.py add code for both GPU and CPU 2020-07-22 11:01:12 +09:00
watershed.py initial commit 2020-07-08 11:31:26 +09:00

README.md

Kindai-OCR

OCR system for recognizing modern Japanese magazines

About

This repo contains an OCR sytem for converting modern Japanese images to text. This is a result of N2I project 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 figures. alt text

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

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, The University of Tokyo and National Institute for Japanese Language and Linguistics, respectively.

Installing Kindai OCR

python==3.7.4
torch==1.4.0
torchvision==0.2.1
opencv-python==3.4.2.17
scikit-image==0.14.2
scipy==1.1.0
Polygon3
pillow==4.3.0

Running Kindai OCR

  • You should first download the pre_trained models and put them into ./pretrain/ folder. VGG model, CRAFT model, OCR model
  • Copy your images into ./data/test/ folder
  • run the following script to recognize images:
    python test.py
  • The recognized text transcription is in ./data/result.xml and the result images are in ./data/result/
  • If you may have to check the path to Japanese font in test.py for correct visualization results.
    fontPIL = '/usr/share/fonts/truetype/fonts-japanese-gothic.ttf' # japanese font
  • using --cuda = True for GPU device and Fasle for CPU device
  • using --canvas_size ot set image size for text line detection
  • An example result from our OCR system

Running Kindai OCR

If you find Kindai OCR useful in your research, please consider citing:
Anh Duc Le, Daichi Mochihashi, Katsuya Masuda, Hideki Mima, and Nam Tuan Ly. 2019. Recognition of Japanese historical text lines by an attention-based encoder-decoder and text line generation. In Proceedings of the 5th International Workshop on Historical Document Imaging and Processing (HIP ’19). Association for Computing Machinery, New York, NY, USA, 37–41. DOI:https://doi.org/10.1145/3352631.3352641

Acknowledgment

We thank The Center for Research and Development of Higher Education, The University of Tokyo, and National Institute for Japanese Language and Linguistics for providing the kindai datasets.

Contact

Dr. Anh Duc Le, email: leducanh841988@gmail.com or anh@ism.ac.jp