JJ's OpenCV - Receipt Challenge - SVCV

In this challenge, the goal is to collect the data from a picture of the receipt. Otherwise, collect data from a picture of a wine bottle.

Assuptions:

• The image is in focus and the text can be read easily.
• The text will be horizontal and in english, left-to-right.
• A logo may be included and, in such case, should be presented seperately. (heheh)
• If analyzed in real-time, multiple frames may provide confirmation of results.

  Sample Reciepts

  Sample Wine

  • "It's a Head Snapper"
    wine_01_front.tsa.txt wine_01_back.tsa.txt

  Tesseract is available as a package in Ubuntu.

  $ lsb_release -a
  No LSB modules are available.
  Distributor ID:	Ubuntu
  Description:	Ubuntu 14.04.1 LTS
  Release:	14.04
  Codename:	trusty
  
  $ sudo apt-get install tesseract-ocr
    ... 
  
  $ tesseract --version
   tesseract 3.03
   leptonica-1.70
    libgif 4.1.6(?) : libjpeg 8d : libpng 1.2.50 : libtiff 4.0.3 : zlib 1.2.8 : webp 0.4.0
  
  $ tesseract receipt_01.jpg stdout > receipt_01.tsa.txt
  
  $ tesseract wine_01_front.jpg stdout > wine_01_front.tsa.txt
  
  $ tesseract wine_01_back.jpg stdout > wine_01_back.tsa.txt
  
  $ tesseract --print-parameters > tesseract_params.txt
  
  $ man tesseract
  

  I suppose I might get better results if I modify some of those parameters. Or, maybe I need to specifically use the english language pack.

  I guess, I thought it would have read the information much better by default...

  Here is how to generate the tesseract and thumbnails

  bash$ for i in *.jpg
  do
  tesseract $i tesseract_$i
  djpeg $i | pnmscale -xsize 200 | cjpeg -opto -progr -qual 75 > small_$i
  echo $i processed
  done
  

  by JJ Stiff - jjstiff arroba hotmail punto com - 31 August 2014