Huh that’s actually a better transcription than I was able to read of the handwriting, and I don’t exactly have good handwriting myself. I just couldn’t see any other reading of validation than ‘voliolation’.
Comment on Why is OCR for handwritten content still that bad?
RobotToaster@mander.xyz 5 days ago
I just asked chatGPT to transcribe it and it said
The handwritten text in the image says:
“Dimer stabilization free energies were also determined from thermodynamic integration (TI, see methods), which provide a direct validation of the MM-GBSA results.”
J. Phys. Chem. B 2018, 122, 7038-7048
There was a post on HN recently about using LLMs for OCR. news.ycombinator.com/item?id=42952605
nimbledaemon@lemmy.dbzer0.com 4 days ago
RobotToaster@mander.xyz 4 days ago
I tried it a while ago out of desperation to read handwriting I couldn’t read at all, that’s why I thought to try it.
hinterlufer@lemmy.world 5 days ago
That’s perfect. Now I’m just wondering why chatGPT is apparently much better in OCR than a dedicated OCR model like EasyOCR or Tesseract.
Btw, Deepseek did a good job but not perfect. I also fed chatGPT a full page of notes and the transcription to markdown worked quite well, although not perfect. However, if I supply the same note as part of a larger pdf, it will refuse to transcribe it, stating that it’s unreadable.
thefactremains@lemmy.world 5 days ago
Because it can fill in gaps where the recognition fails.
executivechimp@discuss.tchncs.de 5 days ago
Which can be problematic. If it makes a mistake and isn’t obviously wrong, that could go unnoticed.
thefactremains@lemmy.world 5 days ago
100% agreed. But it doesn’t change the answer of why they are apparently better than OCR.
homesweethomeMrL@lemmy.world 4 days ago
If I had to guess, I’d say it was the dot paper confusing the OCR reader. I suppose the LLM has some way to cancel out the dots and thereby gets a better scan of it.
cyrano@lemmy.dbzer0.com 5 days ago
Try gemini 2 it seems is pretty good at that as well