i figured it was such a small amount of text that it’d take less time to do manually than it would to track down the original source of the meme (if any exists - it could’ve just been someone messing around with browser devtools or the wikipedia editor preview), but i severely underestimated how hard it is to quickly transcribe text when every second word has been horribly mangled
tbh i’ve never used any form of OCR (outside of the one time i tried out live translation in google translate) and the idea of using it didn’t even cross my mind
It’s named translator (which it is) but it also does OCR and can even live overlay on your camera. And it’s fully offline too after you download the models.
i figured it was such a small amount of text that it’d take less time to do manually than it would to track down the original source of the meme (if any exists - it could’ve just been someone messing around with browser devtools or the wikipedia editor preview), but i severely underestimated how hard it is to quickly transcribe text when every second word has been horribly mangled
I think they were asking why you didn’t just copy it from the picture:
tbh i’ve never used any form of OCR (outside of the one time i tried out live translation in google translate) and the idea of using it didn’t even cross my mind
If you ever do want to try but don’t want to use stupid google services for it then give this a try
https://f-droid.org/packages/dev.davidv.translator
It’s named translator (which it is) but it also does OCR and can even live overlay on your camera. And it’s fully offline too after you download the models.
On the contrary: they asked if there was an original source from which to copy; not why text recognition wasn’t used.
And then the person I’m replying to implied that the picture may be the original source but I understand why you’re saying what you’re saying.