OCR

Text Recognition API

OCR command line utility for macOS 10.15+. Utilizes the VNRecognizeTextRequest API.

print parameters

tesseract --print-parameters

just supply multiple 'parameter'/'config' files.

those files are in 'tessdata' folder. can be found with find / -name tessdata in my case on OSX : /usr/local/Cellar/tesseract-lang/4.0.0/share/tessdata /usr/local/Cellar/tesseract/4.1.1/share/tessdata /usr/local/share/tessdata

tesseract INPUT.JPG OUTPUT_BASE -l nld tsv get.images

get.images gives the intermediate/input image useful for debugging

tesseract INPUT.JPG OUTPUT_BASE -l nld tsv get.images

it writes to: tessinput.tif

more config files:

alto
ambigs.train
api_config
bigram
box.train
box.train.stderr
digits
get.images
hocr
inter
kannada
linebox
logfile
lstm.train
lstmbox
lstmdebug
makebox
pdf
quiet
rebox
strokewidth
tsv
txt
unlv
wordstrbox

tesseract in Python

#!/usr/bin/env python3

from PIL import Image
import pytesseract
print(pytesseract.image_to_data('filelist.txt'))

compressed pdf

  • first compress images using mogrify -quality 40
  • use tesseract with 'filelist.txt'

OCRAD

brew install ocrad
convert 14.16.12.png img.ppm
ocrad img.ppm > output.txt

or

brew install netpbm
pngtopnm filename.png | ocrad

tesseract

brew install tesseract
brew install tesseract-lang

tesseract NL-UtHUA_A356828_000002.jpg outfile -l nld tsv
tesseract NL-UtHUA_A356828_000002.jpg outfile -l nld pdf

other