Has anyone tried to use the image uploader with a phone or tablet to pull the numbers off a card and fill them into a input (store in database)… I was thinking there are lots of cool things we can use this for beyond just capturing credit card info like Uber and other apps do…
You could probably implement the Ocrad.js library via a plug in.
http://antimatter15.com/ocrad.js
I was going to have a go, but then thought I wasn’t ready to get sucked into and distracted by the world of plug-ins just yet!
Maybe something for the Croppie author @mishav to have a go at!
Haha, the thought about doing client-side OCR in a plugin … resembles a nightmare about being trapped in spaghetti, without a fork.
Reading the articles relating to @DaveA’s link:
https://antimatter15.com/project/ocrad-js/
http://antimatter15.com/ocrad.js/demo.html
The two implementations look only vaguely accurate, and “the good one” Tesseract using a neural net sounds too large to run on a client.
Probably have better luck finding an OCR API.
Edit: there is a javascript port of Tesseract, so yes, it may be doable:
http://tesseract.projectnaptha.com/
There’s a basic sample here, include in a HTML element:
https://github.com/naptha/tesseract.js/blob/master/examples/file-input/basic.html
Remove the src script, and instead on page html add the src to the cdn as mentioned here:
You could then put the resulting text to a Bubble value using the Toolbox : )
I haven’t tried it, looking forward to see if you can get it to work.
On the subject of ocr, found this api yesterday! One call later im reading pdf’s and images back into to bubble in less than 8 seconds…
https://ocr.space as usual grab a free api key