Community Wishlist Survey 2019/Multimedia and Commons/Transcription gadget for images with text

Random proposal ►◄ Multimedia and Commons  The survey has concluded. Here are the results!

Transcription gadget for images with text

• Problem: Any text in an image should be transcribed into the description, to make items discoverable
• Who would benefit: Anyone searching commons and people trying to extract info from photographs
• Proposed solution: Gadget to run image through ocr and insert text
• More comments:
• Phabricator tickets:
• Proposer: :JarrahTree (talk) 10:59, 10 November 2018 (UTC)

Discussion

• Available as c:User:Samwilson/GoogleOCR.js. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:51, 17 November 2018 (UTC)

Voting

•  Support Tom Ja (talk) 23:37, 16 November 2018 (UTC)
•  Support Abzeronow (talk) 05:05, 17 November 2018 (UTC)
•  Support Fabiorahamim (talk) 07:05, 17 November 2018 (UTC)
•  Support Gnangarra (talk) 09:19, 17 November 2018 (UTC)
•  Support Liuxinyu970226 (talk) 09:27, 17 November 2018 (UTC)
•  Support ديفيد عادل وهبة خليل 2 (talk) 12:03, 17 November 2018 (UTC)
•  Support Blue Rasberry (talk) 15:39, 17 November 2018 (UTC)
•  Support HLHJ (talk) 06:05, 18 November 2018 (UTC)
•  Support Juandev (talk) 11:21, 18 November 2018 (UTC)
•  Support — Draceane ^talk_contrib. 17:02, 18 November 2018 (UTC)
•  Support Support mainly if the plan is to do this for non-Latin scripts as well. Joalbertine (talk) 17:27, 18 November 2018 (UTC)

  @Joalbertine: If we were to use the Google OCR service, then these languages would be supported — lots of (but not all) non-latin scripts. Sam Wilson 06:08, 20 November 2018 (UTC)

  OMG We must make this happen! --Joalbertine (talk) 09:03, 20 November 2018 (UTC)

•  Support Vulphere 12:11, 20 November 2018 (UTC)
•  Support β₁₆ - (talk) 15:54, 20 November 2018 (UTC)
•  Support Sadads (talk) 22:24, 20 November 2018 (UTC)
•  Support Novak Watchmen (talk) 13:35, 21 November 2018 (UTC)
•  Support I guess we need to implement tools to
  • facilitate people tagging which images have text (AI and/or OCR can be used to pre-guess for a person, person confirms)
  • facilitate people transcribing (computer asks people what language is it; OCR provides an initial version of the text for a person to proofread)
  • facilitate people translating these transcriptions to different languages Gryllida 07:16, 23 November 2018 (UTC)
•  Support Olea (talk) 12:03, 26 November 2018 (UTC)
•  Support Gideoraielgave (talk) 17:22, 26 November 2018 (UTC)
•  Support Filipović Zoran (talk) 19:29, 26 November 2018 (UTC)
•  Support - FlightTime (open channel) 21:48, 26 November 2018 (UTC)
•  Support A garbage person (talk) 17:19, 27 November 2018 (UTC)
•  Support Daniel Case (talk) 23:11, 28 November 2018 (UTC)
•  Support Kpjas (talk) 10:37, 29 November 2018 (UTC)