qurator-spk/neat
1
0
Fork 0
mirror of https://github.com/qurator-spk/neat.git synced 2025-06-09 11:49:54 +02:00
Code Issues Releases Wiki Activity

Update Preprocessing.md

Browse source
This commit is contained in:
Clemens Neudecker 2019-11-20 00:27:10 +01:00 • committed by GitHub
parent 5f6b8bc9c3
commit 9dff8a78ba
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
1 changed files with 8 additions and 4 deletions
Download patch file Download diff file Expand all files Collapse all files

12
docs/Preprocessing.md
View file

@ -3,15 +3,19 @@
The preprocessing pipeline that is developed at the The preprocessing pipeline that is developed at the
[Berlin State Library](http://staatsbibliothek-berlin.de/) [Berlin State Library](http://staatsbibliothek-berlin.de/)
comprises the following steps: comprises the following steps:
- Layout Analysis & Textline Extraction @[sbb_pixelwise_segmentation](https://github.com/qurator-spk/pixelwise_segmentation_SBB)
- OCR & Word Segmentation @[ocrd_tesserocr](https://github.com/OCR-D/ocrd_tesserocr)
- Tokenization
- Named Entity Recognition @[sbb_ner](https://github.com/qurator-spk/sbb_ner)
### Layout Analysis & Textline Extraction ### Layout Analysis & Textline Extraction
Layout Analysis & Textline Extraction @[sbb_pixelwise_segmentation](https://github.com/qurator-spk/pixelwise_segmentation_SBB)
### OCR & Word Segmentation ### OCR & Word Segmentation
OCR is based on [OCR-D](https://github.com/OCR-D)'s [ocrd_tesserocr](https://github.com/OCR-D/ocrd_tesserocr) which requires [Tesseract](https://github.com/tesseract-ocr/tesseract) **>= 4.1.0**. The [Fraktur_5000000](https://ub-backup.bib.uni-mannheim.de/~stweil/ocrd-train/data/Fraktur_5000000/) model, which is trained on [GT4HistOCR](https://github.com/tesseract-ocr/tesstrain/wiki/GT4HistOCR) is used.
corpiu
The [PAGE-XML](https://github.com/PRImA-Research-Lab/PAGE-XML) produced by the [Layout Analysis & Textline Extraction](https://github.com/qurator-spk/neath/blob/master/docs/Preprocessing.md#layout-analysis--textline-extraction) is taken as input, and the output is [PAGE-XML](https://github.com/PRImA-Research-Lab/PAGE-XML) containing the text recognition results with absolute pixel coordinates describing bounding boxes for words.
### Tokenization ### Tokenization
### Named Entity Recognition ### Named Entity Recognition
For Named Entity Recognition, a [BERT-Base](https://github.com/google-research/bert) model was trained. [sbb_ner](https://github.com/qurator-spk/sbb_ner) is using a combination of unsupervised training on a large (~2.3m pages) OCR corpus in combination with supervised training on a small (50k tokens) annotated corpus. Further details are available in the [paper](https://corpora.linguistik.uni-erlangen.de/data/konvens/proceedings/papers/KONVENS2019_paper_4.pdf).