search

OMMR4all / ommr4all-page-segmentation

Other
2 stars 4 forks source link

OMMR4all vs. OCR4all? #2

Closed wrznr closed 4 years ago

wrznr commented 4 years ago

I am confused: The github repo is called ommr4all-page-segmentation but the documentation is for a software called OCR4All Pixel Classifier. How do they relate?

kba commented 4 years ago

Relationship with https://github.com/ocr-d-modul-2-segmentierung/ocrd-pixelclassifier-segmentation would also be interesting.

Gawajn commented 4 years ago

The project was originally used for layout analysis tasks/projects. (one of them is OCR4all). Since the output of the CNN depends on the training material, it can also be used for other purposes. In the Ommr4all project we use it, for example, to segment staves and symbols.

I can not say anything specific to the OCRD project. But the original repo of the project is located at https://gitlab2.informatik.uni-wuerzburg.de/ls6/ocr4all-pixel-classifier