Natanox@discuss.tchncs.de to linuxmemes@lemmy.worldEnglish · 21 天前Easy, it's uuuuuuuuh…discuss.tchncs.deexternal-linkmessage-square58fedilinkarrow-up1494arrow-down14
arrow-up1490arrow-down1external-linkEasy, it's uuuuuuuuh…discuss.tchncs.deNatanox@discuss.tchncs.de to linuxmemes@lemmy.worldEnglish · 21 天前message-square58fedilink
minus-squarewander1236@sh.itjust.workslinkfedilinkarrow-up33arrow-down1·21 天前OCR of fonts used to be a solved problem, but now we have AI, which can sort of do it sometimes
minus-squareerrer@lemmy.worldlinkfedilinkEnglisharrow-up13arrow-down1·21 天前Why be boring and do it right when you can vibe some letters instead?
minus-squaremorrowind@lemmy.mllinkfedilinkarrow-up12arrow-down1·21 天前OCR was AI. Anyway today’s models are measurably better especially when you go beyond simple text on a clean page.
minus-squarevivendi@programming.devlinkfedilinkEnglisharrow-up5arrow-down1·21 天前Any good OCR model also uses “AI” And LLMs are usually really good at detecting text Source: Had to OCR a quite a few ancient university papers
OCR of fonts used to be a solved problem, but now we have AI, which can sort of do it sometimes
Why be boring and do it right when you can vibe some letters instead?
OCR was AI.
Anyway today’s models are measurably better especially when you go beyond simple text on a clean page.
Any good OCR model also uses “AI”
And LLMs are usually really good at detecting text
Source: Had to OCR a quite a few ancient university papers