The workflow is going to be adapted to allow the use of OCRed text as input. The OCR engine is going to be ​OCRopus.

​Tutorial video and ​other videos

The documents of the previous workflows were assessed in terms of how well they might perform being OCRed.

• Easy
  • Bernstein, 1897 (Fraktur)
  • ​Berzelius 1819
  • ​Ampère 1844
• Medium
  • ​Vitruvius 1511
  • ​Cataneo 1600
  • ​Aristoteles 1547 (italics)
  • ​Archimedes 1565 (many pictures (problematic?))
  • ​Cataneo 1572 (bad printing)
  • ​Viviani 1659
  • ​Bianconi 1746
  • ​Zanotti 1752
  • ​Bion 1765 (blackletter)
  • ​Vitruvius 1800 (clear, but weak printing, can see printing of the other side)
  • ​Gallaccini 1767 (small font, but quite clear)
  • ​Angeli 1668 (contains italics, otherwise very clear)
  • ​Trigault 1639
  • ​Bernoulli 1738
• Hard
  • ​Vitruvius 1544
  • ​Vitruvius 1757 (Mixed blackletter and antiqua)
  • ​Zonca 1656 (bad printing)
  • ​Bacon 1670 (contains italics, bad printing)
  • ​Clavius 1606
  • ​Barrow 1674 (bad printing, italics and Greek)
  • ​Gravesande 1721
  • ​Vitruvius 1618 (Thesaurus has columns)
  • ​Mersenne 1635 (Microfilm)
  • ​Aristoteles 1548 (contains Greek)
  • ​Vitruvius 1556 (extremly small font)
  • ​Aristoteles 1585 (bad printing)
  • ​Specklin 1599 (bad printing, blackletter)
  • ​Biancani 1635 (very small font)
  • ​Vitruvius 1567 (very small font, mixed italics and upright)
  • ​Archimedes 1565 (mixed italics and upright)

Command overview

The following commands (taken from above video) allow the recognition of English text:

1. ocropus-binarize 035.jpg
2. ocropus-pseg book/????.png
3. ocropus-lattices -m OCRopus/ocropy/2m2-reject.cmodel book/0001/??????.png
4. ocropus-align -l OCRopus/ocropy/data/default.fst book/0001/??????.fst
5. ocropus-hocr book/