My recollection of OCR – optical character recognition – from years ago was it was expensive, fiddly, and not very good. Nowadays a free product like Google Docs has it built in.
I saw this post by Chris Aldrich: Typecasting with IndieWeb Building Blocks and one of his steps is to use Google Docs to recognize the text from an image. That process requires uploading the image and then doing “Open as… Google Docs” … I was curious if it was possible from handwritten text. And so I created this and uploaded it:
INSPIRED BY BOFFOSOCKO dot com let’s test the ability of Google docs to recognize text in AN IMAGE, MAY be MAybe to blog it
… and that matches the text I wrote out by hand awfully well. It has an extra “maybe maybe” couplet, perhaps attempting to provide me options for what I may have meant.
But a handwritten workflow for posting to a blog, ably and fully explored on the indieweb wiki page for handwriting is within reach and a possibility. Super fun stuff, here at IndieWebCamp San Diego Day 2.
The prevalence and availability of image to text feels very much like magic for me, still.
