4 ms·

What do folks think about these document types as a corpus for comparing tools? It's missing images and handwriting samples, but those types of documents might

by iisan7 4y ago

What do folks think about these document types as a corpus for comparing tools? It's missing images and handwriting samples, but those types of documents might just be too variable to make conclusions about. I remember Baidu's OCR giving excellent English results, but it looks like their API is deprecated now. Out of curiousity, I ran these samples through easyOCR by JaidedAI. Results at https://pastebin.com/RjzVd5Sf https://pastebin.com/RjzVd5Sf.

sgc 4y ago
I OCR books, so they are not a good sample. I would want to compare at least 10 pages per sample, with more typical problems such as skewed, rounded pages from photos, artifacts, damaged source pages (tears and creases) etc. They do reproduce some problems with changing fonts and layout, but a big piece of the puzzle is custom dictionaries and layout training. It's fine for a once over, but not a deep dive.
- llanowarelves 4y ago
  What are your favorites overall for book scanning? I'm building a DIY scanner and have only briefly considered the actual software I'll use in the processing pipeline. FOSS or API tooling preferred unless proprietary packages are significantly better.
  - sgc 4y ago
    ABBYY has been much better than anything else due to its abilities to fine tune layout, recognition, and export parameters. I follow threads like this, always looking for the latest and greatest, but nothing else is worth the time for a smaller organization. We scan several languages. If you are English only, I can't answer about recognition since I don't look. But layout and export are mission critical, and worth a few hundred bucks if you can afford it.
    - llanowarelves 4y ago
      Thanks for the response. Which ABBYY product(s) is this? I'm a little confused by their website, seems like they offer quite a lot of combinations of things.
      - sgc 4y ago
        We use an old stand-alone version of FineReader. I cannot tell you if their current offerings are worth the price.

by iisan7 4y ago

sgc 4y ago

I OCR books, so they are not a good sample. I would want to compare at least 10 pages per sample, with more typical problems such as skewed, rounded pages from photos, artifacts, damaged source pages (tears and creases) etc. They do reproduce some problems with changing fonts and layout, but a big piece of the puzzle is custom dictionaries and layout training. It's fine for a once over, but not a deep dive.

llanowarelves 4y ago
What are your favorites overall for book scanning? I'm building a DIY scanner and have only briefly considered the actual software I'll use in the processing pipeline. FOSS or API tooling preferred unless proprietary packages are significantly better.
- sgc 4y ago
  ABBYY has been much better than anything else due to its abilities to fine tune layout, recognition, and export parameters. I follow threads like this, always looking for the latest and greatest, but nothing else is worth the time for a smaller organization. We scan several languages. If you are English only, I can't answer about recognition since I don't look. But layout and export are mission critical, and worth a few hundred bucks if you can afford it.
  - llanowarelves 4y ago
    Thanks for the response. Which ABBYY product(s) is this? I'm a little confused by their website, seems like they offer quite a lot of combinations of things.
    - sgc 4y ago
      We use an old stand-alone version of FineReader. I cannot tell you if their current offerings are worth the price.