OCR Output Postprocessing Pipeline
split · stats · extract · classify · aggregate
Post-processing of OCR output — ALTO XML, and also PAGE XML, hOCR, PDF, office and text files — into extracted text and a scored line-quality table, plus per-page ALTO.
Deserialisation of ALTO and every other OCR or text-bearing input, and OCR quality control. The fan-out point of the whole pipeline.
- python3.11
- 🤗 HFfasttext-langID
- 🤗 HFQwen2.5-0.5B
- vendoredalto-tools
- SSH OMPYParYU
- licenseMIT
- funded byATRIUM
Where this sits in the pipeline
- page-classification
- →
- ocr-postprocess
- →
- translator
- →
- nlp-enrich
- →
- keyword-extract
- →
- digital-convert