Conversation
Notices
-
any actually blind people here on fedi who can tell me if they use software to describe images / OCR stuff?
-
@piggo that already works, but especially OCR has been working for decades now, so it seems strange to me that blind people wouldn't use it, but I don't know anything about it.
-
@lain saw them talk of using AI for that recently
-
@lain @piggo OCR works but 2023 AI works way better wirth suboptimal inputs, it's not even comparable
t. tried running OCR on 1940s phonebook scans yesterday and it wasn't great
-
@eal @piggo yeah, i mean more for stuff like screenshots
-
@eal @piggo i wouldn't think they'd know it, i'd just think that hitting 'ocr and read' on any random image would be normal, and the results should be good for things like screenshots of text.
-
@lain @piggo how would a blind person know if it's a screenshot before using a voice tool on it?
-
@eal @piggo i also recently used chatgpt to transcribe a horrible video screenshot of a 50 year old machine-typed recipe with coffee stains and everything, worked really well when 'normal' OCR failed.
-
@lain Maybe ask @marco
-
@pony @piggo @eal chatgpt can do it (via api, too), i tried the local variants (bakllava, llava) but they aren't good. Apple released something a month ago in the same space but i haven't tried it.
-
@eal @lain @piggo is there some ready made solution I can use? I really wanted to do something with my journals but like it’s not that easy to read even for me