[dead]
How we made our OCR code more accurate
11–20 of 50 posts
Re: How we made our OCR code more accurate
#12Anything that mentions tesseract is about 10 years out of date at this point.
I just built a pipeline with tesseract last year. What's better that is open source and runnable locally?
VLLM hallucination is a blocker for my use case.
Re: How we made our OCR code more accurate
#13Neat article, but I feel like I have no idea why they're doing this! Is transcribing code from images really such a big use case?
The product appears to be similar to Microsoft's embattled Recall feature. In order to remember your digital life it takes frequent screenshots.
Re: How we made our OCR code more accurate
#14Anything that mentions tesseract is about 10 years out of date at this point.
I just built a pipeline with tesseract last year. What's better that is open source and runnable locally? VLLM hallucination is a blocker for my use case.
How is a hallucination worse than a Tesseract error?
Re: How we made our OCR code more accurate
#15Earlier quoted context omitted.
I just built a pipeline with tesseract last year. What's better that is open source and runnable locally? VLLM hallucination is a blocker for my use case.
How is a hallucination worse than a Tesseract error?
Hallucinations are hard to detect unless you are a subject-matter expert. I don't have direct experience with Tesseract error detection.
Re: How we made our OCR code more accurate
#16Anything that mentions tesseract is about 10 years out of date at this point.
Quite simply, you’re completely wrong. Modern tesseract versions include a modern LSTM AI. It can very affordably be deployed on CPU, yet its performance is competitive with much more expensive large GPU-based models. Especially if you handle a high volume of scans, chances are that tesseract will have the best bang per buck.
Re: How we made our OCR code more accurate
#17Tesseract OCR was created by digital (DEC) in 19_8_5 (yes, 40 not four YEARs ago). Now go back and read the article and ROFL with me.
Re: How we made our OCR code more accurate
#18Earlier quoted context omitted.
I just built a pipeline with tesseract last year. What's better that is open source and runnable locally? VLLM hallucination is a blocker for my use case.
How is a hallucination worse than a Tesseract error?
Because the VLM doesn't know it hallucinated. When you get a Tesseract error you can flag the OCR job for manual review.
Re: How we made our OCR code more accurate
#19Earlier quoted context omitted.
I just built a pipeline with tesseract last year. What's better that is open source and runnable locally? VLLM hallucination is a blocker for my use case.
How is a hallucination worse than a Tesseract error?
It could hallucinate obscene language, something which is less likely with classic OCR.
Re: How we made our OCR code more accurate
#20Tesseract OCR was created by digital (DEC) in 19_8_5 (yes, 40 not four YEARs ago). Now go back and read the article and ROFL with me.
The original tesseract OCR has no neural nets. It bare little resemblance to the modern version.