8 ms·

Appears to be a nice wrapper around Tesseract: https://github.com/tesseract-ocr/tessdata https://github.com/tesseract-ocr/tessdata https://en.wikipedia.org/wi

Appears to be a nice wrapper around Tesseract: https://github.com/tesseract-ocr/tessdata https://github.com/tesseract-ocr/tessdata https://en.wikipedia.org/wiki/Tesseract_(software) https://en.wikipedia.org/wiki/Tesseract_(software) The demo of course works perfectly on a Mac as this is already built into Ventura. If you haven't experienced it yet ye olde ctrl-f now seamlessly sneaks a peak into images on the page for example, surprisingly useful. In November 2020, Brewster Kahle from the Internet Archive praised Tesseract saying: Tesseract has made a major step forward in the last few years. When we last evaluated the accuracy it was not as good as the proprietary OCR, but that has changed– we have done evaluations and it is just as good, and can get better for our application because of its new architecture. Anybody have an up to date breakdown of available OCR solutions?

Icko 4y ago
Last I compared them, (1-2 years ago), Google OCR was much much better and supported more languages than tesseract. There was also an OCR in openCV, which was slightly better than tesseract, but not good enough to be useful.
- rafram 4y ago
  I’m not aware of any separate OCR in OpenCV. Some builds include an interface to Tesseract, which might be what you’re thinking of. Tesseract certainly benefits from preprocessing (conversion to grayscale, posterization) with OpenCV.
  - Icko 4y ago
    There was "EAST OCR" detector, which is basically someone put a deep learning model in openCV somehow. https://www.folio3.ai/blog/text-detection-by-using-opencv-and-east/ https://www.folio3.ai/blog/text-detection-by-using-opencv-an...
- lasagna_coder 4y ago
  What are these projects are you referring to? AFAIK Tesseract is sponsored by Google, from what I understand it is state of the art, ie it is Google OCR. Searching for OCR with OpenCV only reveals using OpenCV with Tesseract, not rolling its own OCR, OpenCV being used to preprocess images to optimise them for Tesseract. Maybe I'm missing something, so I'm interested if you can point me in the right direction.
  - spi 4y ago
    Google OCR is definitely not the same as Tesseract, although it's true that Tesseract is maintained by Google. Google OCR has definitely much higher accuracy and is significantly faster (basically always taking 1s for inference, while Tesseract can easily take 10s or more for dense pages). Source: I work in developing a competing OCR service and we keep an eye on competition (e.g. aside from Google, solutions by Azure, Amazon, Abbyy, Nuance, Cloudmersive, etc., as well as our internal product of course, which is not available externally), and they are (almost) all significantly better on Tesseract. The only domain where Tesseract is competitive is for perfect "black text on white paper", it gives pretty poor performance when dealing with colored, distorted text, or even strong page structure effects (tables, etc.). When I say "pretty poor" I mean: "with respect to the state-of-the-art", of course it's still enormously better than what was the state-of-the-art before deep learning came into the picture, roughly a decade ago. And for things like "search contents of a book" it's basically perfect already.
    - recuter 4y ago
      > Source: I work in developing a competing OCR service and we keep an eye on competition (e.g. aside from Google, solutions by Azure, Amazon, Abbyy, Nuance, Cloudmersive, etc., as well as our internal product of course, which is not available externally), and they are (almost) all significantly better on Tesseract. Great. How do you quantify it and keep track? Is there an industry standard benchmark? Would you consider sharing a backblaze type analysis (they track consumer HD performance and blogging about it got them a lot of attention and customers)?
      - deleted 4y ago
- holbue 4y ago
  I agree, there are way better cloud based and proprietary OCR solutions out there. But Tesseract still seems to deliver the best results among the FOSS tools, doesn't it? Back in the days, Cuneiform got close to Tesseract's performance, but AFAIK it wasn't developed further... Does anyone else know other promising open-source OCR engines?
nickserv 4y ago
There's also DocTR which can do text detection and extraction out of the box. It's command line driven but can display the detected text as an overlay of the document. https://github.com/mindee/doctr https://github.com/mindee/doctr
jibbers 4y ago
And for older Mac OS's T-Rex https://trex.ameba.co https://trex.ameba.co
captnswing 4y ago
On a Mac, for ad-hoc OCR, I use the immensely useful CleanShot X https://cleanshot.com/ https://cleanshot.com/ (which is well worth paying for). Among many other things, it offes OCR of any region on the screen for larger-scale OCR processing of pdfs and other files, I love how s3-ocr https://simonwillison.net/2022/Jun/30/s3-ocr/ https://simonwillison.net/2022/Jun/30/s3-ocr/ makes working with AWS Textract OCR more accessible (though, somehow, Textract refuses to fully OCR larger pdfs I possess..)
- kube-system 4y ago
  On the latest MacOS, OCR happens automatically in any screenshot, or any image you open in Preview. Try Command+Shift+4, grab part of the screen, click the pop-up, and just select text.
ce4 4y ago
I use https://kebekus.gitlab.io/scantools https://kebekus.gitlab.io/scantools for scanning, it builds on top of tesseract and works great for pdf enhancements
- rjzzleep 4y ago
  You might be interested in https://github.com/ocrmypdf/OCRmyPDF https://github.com/ocrmypdf/OCRmyPDF then. It does quite some preprocessing on the PDF pages before passing it on to tesseract.
  - angrygoat 4y ago
    I've found ocrmypdf to be excellent: the only issue I've had is with PDFs with differing page sizes; it seems to scale everything up to the size of the largest page, which can be a bit of a pain.
bjacobt 4y ago
I’ve had good results from paddle ocr. https://github.com/PaddlePaddle/PaddleOCR https://github.com/PaddlePaddle/PaddleOCR
IceHegel 4y ago
In 2019 I was working on a project that involved OCRing millions of scanned historical documents. I evaluated Google, Azure, Amazon, Adobe, ABBYY, and Tesseract somewhat rigorously. Google's was by far the best, especially for obscured or malformed characters. Azure was second and I ended up merging the results from both. For my use case (in Spring 2019) Tesseract was not very accurate and struggled with slanted text especially. Hopefully that has changed.

Icko 4y ago

Last I compared them, (1-2 years ago), Google OCR was much much better and supported more languages than tesseract. There was also an OCR in openCV, which was slightly better than tesseract, but not good enough to be useful.

rafram 4y ago
I’m not aware of any separate OCR in OpenCV. Some builds include an interface to Tesseract, which might be what you’re thinking of. Tesseract certainly benefits from preprocessing (conversion to grayscale, posterization) with OpenCV.
- Icko 4y ago
  There was "EAST OCR" detector, which is basically someone put a deep learning model in openCV somehow. https://www.folio3.ai/blog/text-detection-by-using-opencv-and-east/ https://www.folio3.ai/blog/text-detection-by-using-opencv-an...
lasagna_coder 4y ago
What are these projects are you referring to? AFAIK Tesseract is sponsored by Google, from what I understand it is state of the art, ie it is Google OCR. Searching for OCR with OpenCV only reveals using OpenCV with Tesseract, not rolling its own OCR, OpenCV being used to preprocess images to optimise them for Tesseract. Maybe I'm missing something, so I'm interested if you can point me in the right direction.
- spi 4y ago
  Google OCR is definitely not the same as Tesseract, although it's true that Tesseract is maintained by Google. Google OCR has definitely much higher accuracy and is significantly faster (basically always taking 1s for inference, while Tesseract can easily take 10s or more for dense pages). Source: I work in developing a competing OCR service and we keep an eye on competition (e.g. aside from Google, solutions by Azure, Amazon, Abbyy, Nuance, Cloudmersive, etc., as well as our internal product of course, which is not available externally), and they are (almost) all significantly better on Tesseract. The only domain where Tesseract is competitive is for perfect "black text on white paper", it gives pretty poor performance when dealing with colored, distorted text, or even strong page structure effects (tables, etc.). When I say "pretty poor" I mean: "with respect to the state-of-the-art", of course it's still enormously better than what was the state-of-the-art before deep learning came into the picture, roughly a decade ago. And for things like "search contents of a book" it's basically perfect already.
  - recuter 4y ago
    > Source: I work in developing a competing OCR service and we keep an eye on competition (e.g. aside from Google, solutions by Azure, Amazon, Abbyy, Nuance, Cloudmersive, etc., as well as our internal product of course, which is not available externally), and they are (almost) all significantly better on Tesseract. Great. How do you quantify it and keep track? Is there an industry standard benchmark? Would you consider sharing a backblaze type analysis (they track consumer HD performance and blogging about it got them a lot of attention and customers)?
    - deleted 4y ago
holbue 4y ago
I agree, there are way better cloud based and proprietary OCR solutions out there. But Tesseract still seems to deliver the best results among the FOSS tools, doesn't it? Back in the days, Cuneiform got close to Tesseract's performance, but AFAIK it wasn't developed further... Does anyone else know other promising open-source OCR engines?

nickserv 4y ago

There's also DocTR which can do text detection and extraction out of the box. It's command line driven but can display the detected text as an overlay of the document. https://github.com/mindee/doctr https://github.com/mindee/doctr

jibbers 4y ago

And for older Mac OS's T-Rex https://trex.ameba.co https://trex.ameba.co

captnswing 4y ago

On a Mac, for ad-hoc OCR, I use the immensely useful CleanShot X https://cleanshot.com/ https://cleanshot.com/ (which is well worth paying for). Among many other things, it offes OCR of any region on the screen for larger-scale OCR processing of pdfs and other files, I love how s3-ocr https://simonwillison.net/2022/Jun/30/s3-ocr/ https://simonwillison.net/2022/Jun/30/s3-ocr/ makes working with AWS Textract OCR more accessible (though, somehow, Textract refuses to fully OCR larger pdfs I possess..)

kube-system 4y ago
On the latest MacOS, OCR happens automatically in any screenshot, or any image you open in Preview. Try Command+Shift+4, grab part of the screen, click the pop-up, and just select text.

ce4 4y ago

I use https://kebekus.gitlab.io/scantools https://kebekus.gitlab.io/scantools for scanning, it builds on top of tesseract and works great for pdf enhancements

rjzzleep 4y ago
You might be interested in https://github.com/ocrmypdf/OCRmyPDF https://github.com/ocrmypdf/OCRmyPDF then. It does quite some preprocessing on the PDF pages before passing it on to tesseract.
- angrygoat 4y ago
  I've found ocrmypdf to be excellent: the only issue I've had is with PDFs with differing page sizes; it seems to scale everything up to the size of the largest page, which can be a bit of a pain.

bjacobt 4y ago

I’ve had good results from paddle ocr. https://github.com/PaddlePaddle/PaddleOCR https://github.com/PaddlePaddle/PaddleOCR

IceHegel 4y ago

In 2019 I was working on a project that involved OCRing millions of scanned historical documents. I evaluated Google, Azure, Amazon, Adobe, ABBYY, and Tesseract somewhat rigorously. Google's was by far the best, especially for obscured or malformed characters. Azure was second and I ended up merging the results from both. For my use case (in Spring 2019) Tesseract was not very accurate and struggled with slanted text especially. Hopefully that has changed.