Talk:Optical character recognition/Archives/2020

Replace misnomer?

Optical character recognition is a misnomer. Even in the past, software did not recognize characters, but glyphs. Modern software which use artificial neural networks no longer focus on characters or glyphs but analyze whole lines of text. Therefore text recognition might be a better title for this article. --Stefan Weil (talk) 09:01, 17 January 2018 (UTC)

A few different ways to look at it. One is that OCR is more a matter of being the common name for it without necessarily being the best technical description. Another is that OCR often includes the additional task of interpreting alpha-numeric characters from glyphs. North8000 (talk) 13:57, 17 January 2018 (UTC).

@North8000 and Stefan Weil: "Text recognition" appears to be a common synonym of "optical character recognition," so it should at least be mentioned in the lead section. Jarble (talk) 19:58, 21 February 2020 (UTC)

I was concerned that text recognition might be a broader topic than OCR. But I googled it and looked at the firs ~50 hits and it appears to nearly always mean OCR. North8000 (talk) 21:35, 21 February 2020 (UTC)

But OCR can include non-text characters. North8000 (talk) 12:17, 22 February 2020 (UTC)

There are no non-text characters as long as you define text to be a collection of characters. And that's what an OCR software usually does. --Stefan Weil (talk) 13:00, 22 February 2020 (UTC)

Some confusing details here

@MikeTernoey: This edit describes OCR for "CAD images in a database that are appropriate to the vehicle design," though I don't know what it specifically refers to. Does this article include any sources that should be cited here? Jarble (talk) 19:45, 21 February 2020 (UTC)

IMO it's not clear or informative, not sourced, and possibly not be about OCR. Should probably be taken out. North8000 (talk) 19:50, 21 February 2020 (UTC)