Batch OCR with Open Source Tools

1
Batch OCR withOpen Source Tools

• Jonathan Brinley
• Adelie Design
• (ex-Ball State University)

2

• http//whatever.scalzi.com/2006/09/13/clearly-you-
  people-thought-i-was-kidding/

3

• Tesseract
• http//code.google.com/p/tesseract-ocr/
• OCRopus
• http//code.google.com/p/ocropus/

4
How to OCR an Image

• ocroscript recognize /path/to/file.png gt
  /path/to/output.html

5
hOCR

• ltbodygt
• ltdiv class"ocr_page" title"bbox 0 0 2548 3300
  image /path/to/scanned/image.png"gt
• ltspan class"ocr_line" title"bbox 659 143 863
  177"gtSome Textlt/spangt
• ltspan class"ocr_line" title"bbox 723 275 916
  324"gtMore Textlt/spangt
• lt/divgt
• lt/bodygt

6

• http//www.brainofshawn.com/2007/08/05/givin-scalz
  i-a-hand/

7
HocrConverter.py

• from HocrConverter import HocrConverter
• hocr HocrConverter("myHocrFile.html")
• hocr.to_text("output.txt")
• hocr.to_pdf("myImageFile.png", "output.pdf")

8
Learn More or Get the Code

• http//xplus3.net/2009/04/02/convert-hocr-to-pdf/
• jonathanbrinley_at_gmail.com