Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The writing is on the wall for handwriting recognition

newsletter.dancohen.org

Post reply on HN
Print view View original

41–50 of 116 posts

Re: The writing is on the wall for handwriting recognition

#41

Post by GaggiX »

lccerina wrote: post #35
coolness wrote: post #2

Great post and amazing progress in this field! However, I have to wonder if some of these letters were part of the training data for Gemini, since they are well-known and someone has probably already done the painstaking work of transcribing them...

Most likely, and probably inferring the structure on texts with "similar" writing forms. Tried with my handwriting (in italian) and the performance wasn't that stellar. More annoyingly, it is still a LLM and not a "pure" OCR, so some sentences were partially rephrased with different words than the one in the text. This is crucially problematic if they would be used to transcribe historical documents

Are you sure to have used the Gemini 3.0 pro model? Maybe try increasing the media resolution on the AI studio if the text is small

Re: The writing is on the wall for handwriting recognition

#42

Post by embedding-shape »

sph wrote: post #25

Any self-hosted open source solution? I would like to digitize my paper notebooks but I do not want to use anything proprietary or that uses external services. What is the state of the art on the FOSS side? Ideally something that I can train with my own handwriting. I had a look at Tesseract, wondering if there’s anything better out there.

Try various downloadable weights that has Vision, they're all good at different examples, running multiple ones and then finally something to aggregate/figure out the right one usually does the trick. Some recent ones to keep in the list: ministral-3-14b-reasoning, qwen3-vl-30b, magistral-small-2509, gemma-3-27b

Personally I found magistral-small-2509 to be overall most accurate, but it completely fails on some samples, while qwen3-vl-30b doesn't struggle at all with those same samples. So seems training data is really uneven depending on what exactly you're trying to OCR.

And the trade-off of course is that these are LLMs so not exactly lightweight nor fast on consumer hardware, but at least with the approach of using multiple you greatly increase the accuracy.

Re: The writing is on the wall for handwriting recognition

#43

Post by seethishat »

notimetorelax wrote: post #38
benterix wrote: post #28

> If AI can diminish some of the monotony of research, perhaps we can spend more time thinking, writing, playing piano, and taking walks — with other people. Whenever any progress is made, this is the logical conclusion. And yet, those who decide about how your time is being used, have an opposing view.

I feel that we’re reaching a limit to our context switching. Any further process improvements or optimizations will be bottlenecked on humans. And I don’t think AI will help here as jobs will account for that and we’ll have to do context switching on even broader and more complex scopes.

I think the limit has been exceeded. That's the primary reason everything sort of sucks now. There is no time to slow down and do things right (or better).

IMO, cyber security, for example, will have to become a government mandate with real penalties for non-compliance (like seat belts in cars were mandated) in order to force organizations to slow down, and make sure systems are built carefully and as correctly as possible to protect data.

This is in conflict with the hurtling pace of garbage in/garbage out AI generated stuff we see today.

Re: The writing is on the wall for handwriting recognition

#45

Post by myth_drannon »

__alexs wrote: post #7

Call me when it can do Russian Cursive.

Right, it can do modern writing but anything older than a century ( church records and census)and it produces garbage. Yandex Archives figured that out and have CER in a single digit but they have the resources to collect immense data for training. I'm slowly building a dataset for finetuning TROCR model and the best it can do is CER 18% ... which is sort of readable.

Re: The writing is on the wall for handwriting recognition

#47

Post by butlike »

lccerina wrote: post #35
coolness wrote: post #2

Great post and amazing progress in this field! However, I have to wonder if some of these letters were part of the training data for Gemini, since they are well-known and someone has probably already done the painstaking work of transcribing them...

Most likely, and probably inferring the structure on texts with "similar" writing forms. Tried with my handwriting (in italian) and the performance wasn't that stellar. More annoyingly, it is still a LLM and not a "pure" OCR, so some sentences were partially rephrased with different words than the one in the text. This is crucially problematic if they would be used to transcribe historical documents

So it doesn't work is what you're saying, right?

Re: The writing is on the wall for handwriting recognition

#50

Post by sph »

nikanj wrote: post #17

The writing is on the wall for handwriting. Zoomers use speech recognition or touchscreen keyboards, millennials use keyboards. Boomers use pens

Silly comment. Handwriting is proven to be correlated with much better memory retention, which ultimately means much greater degree of association with existing memories and the creation of novel ideas.

"The comparison between handwriting and typing reveals important differences in their neural and cognitive impacts. Handwriting activates a broader network of brain regions involved in motor, sensory, and cognitive processing, contributing to deeper learning, enhanced memory retention, and more effective engagement with written material. Typing, while more efficient and automated, engages fewer neural circuits, resulting in more passive cognitive engagement. These findings suggest that despite the advantages of typing in terms of speed and convenience, handwriting remains an important tool for learning and memory retention, particularly in educational contexts."

https://pmc.ncbi.nlm.nih.gov/articles/PMC11943480/

You are literally handicapping yourself by not thinking with pen and paper, or keeping paper notes.

The future is handwriting with painless digitization for searchability, until we invent a better input device for text that leverages our motor-memory facilities in the brain.

Post reply on HN