Difference between revisions of "OCR/Roadmap"

From Open Food Facts wiki
Jump to navigation Jump to search
Line 14: Line 14:
 
* Use the right standard dict for each language
 
* Use the right standard dict for each language
 
* Integrate custom lists from Global Ingredients Taxonomy  
 
* Integrate custom lists from Global Ingredients Taxonomy  
**Create a golden set
+
* USDA UNII list of ingredients (will also work for Open Beauty Facts)
*** e.g.  someproduct.jpg -> ingredients image
 
*** someproduct.golden -> ingredients text
 
*** then we create a script that runs the OCR through the images, compare with the golden text, and report some accuracy measures
 
*** Draft Script: https://lite6.framapad.org/p/OFF_OCR_Script
 
 
* Integrate custom lists from the live instances; language per language.
 
* Integrate custom lists from the live instances; language per language.
 
** http://de.openfoodfacts.org/zutaten
 
** http://de.openfoodfacts.org/zutaten
 
** http://uk.openfoodfacts.org/ingredients + http://us.openfoodfacts.org/ingredients
 
** http://uk.openfoodfacts.org/ingredients + http://us.openfoodfacts.org/ingredients
 
**http:// fr.openfoodfacts.org/ingredients
 
**http:// fr.openfoodfacts.org/ingredients
** USDA UNII list of ingredients (will also work for Open Beauty Facts)
 
* Process all images and make products searchable, even if not filled yet
 
 
===Dictionaries ===
 
===Dictionaries ===
 
* https://openfoodfacts.slack.com/files/teolemon/F08FC3T6V/deu.user-words
 
* https://openfoodfacts.slack.com/files/teolemon/F08FC3T6V/deu.user-words
Line 31: Line 25:
 
* https://openfoodfacts.slack.com/files/teolemon/F08FBQ45D/nld.user-words
 
* https://openfoodfacts.slack.com/files/teolemon/F08FBQ45D/nld.user-words
 
* https://openfoodfacts.slack.com/files/teolemon/F08FBQ45V/spa.user-words
 
* https://openfoodfacts.slack.com/files/teolemon/F08FBQ45V/spa.user-words
 +
 +
=== Testing ===
 +
**Create a golden set
 +
*** e.g.  someproduct.jpg -> ingredients image
 +
*** someproduct.golden -> ingredients text
 +
*** then we create a script that runs the OCR through the images, compare with the golden text, and report some accuracy measures
 +
*** Draft Script: https://lite6.framapad.org/p/OFF_OCR_Script
 +
 +
 +
 +
 +
* Process all images and make products searchable, even if not filled yet
  
 
==  Long-term goals ==
 
==  Long-term goals ==

Revision as of 09:36, 18 August 2015

Currently, all products are edited manually. This project is about automatic or semi-automatic detection of a number of things using OCR and Computer vision.

Current state

  • OCR extraction of Ingredients using Tesseract 2 (production) and 3 (.net)
  • Uses the French dictionary for all languages
-- /home/off-fr/cgi# grep get_ocr *
Ingredients.pm:use Image::OCR::Tesseract 'get_ocr';
Ingredients.pm: $text =  decode utf8=>get_ocr($image,undef,'fra');

Short term goals

Dictionaries

Testing

    • Create a golden set
      • e.g. someproduct.jpg -> ingredients image
      • someproduct.golden -> ingredients text
      • then we create a script that runs the OCR through the images, compare with the golden text, and report some accuracy measures
      • Draft Script: https://lite6.framapad.org/p/OFF_OCR_Script



  • Process all images and make products searchable, even if not filled yet

Long-term goals

  • Get dictionaries translations from Wikidata
  • Investigate Ocropus for complex layout extractions
  • Investigate Open CV for detection of patterns, logos…

Targets

  • Logos of brands (Getting them from POD ?)
  • Logos of Labels (standardized)
  • Text (distorted - bottle case, diagonally - with low light, bright light)
  • Standardized layouts (US Nutrition labels)
    • Store in separate image for further reference
  • Standardized text (quantities, EU Packaging codes)
    • Store in separate image for further reference
  • Barcodes (extraction in uploaded images)
    • Store in separate image for further reference
  • Image orientation: check that the text is properly oriented to guess if the image is properly oriented.
  • Deep Learning
    • Product photo on packaging - guess category based on product picture
    • Container: guess whether it's a bottle, cardboard…

Extracting areas is already great work: if we can extract logos or patterns, it will be faster for humans to double check and turn that into text.