!Tesseract

I’ve tried to run !Tesseract on a Pi3 B+, but typing tesseract on its own in a TaskWindow, after 5 minutes nothing has happened (should generate the help text). Does it just need more resources than a Pi3 can provide?

Not sure, maybe I not use the latest one…
For Pi3B+ “my” tesseract inside !Tesseract needs to patch with PatchSWP.

You’re right—I’ve just tried patching tesseract (the !PatchSWP program was already on RO Direct in $.Utilities.Patches) and it’s now fine. It’s certainly faster running on this Pi3 than it is running on my Linux laptop if I have RPCEmu running at the same time.

There’s an updated version of Tesseract OCR Engine v3.05.02 available here. It now supports about 100 languages.

It might be an idea if you could port Tesseract 4—I think officially it’s a “beta” but it’s much more accurate than Tesseract 3.

It might be an idea if you could port Tesseract 4—I think officially it’s a “beta” but it’s much more accurate than Tesseract 3.

I’ve done a port of 5.00alpha and it is much more accurate. Once the next version of GCCSDK is released I can make it available for download. It only builds with gcc 8.

There’s an updated version of Tesseract OCR Engine v5.0.0 available here.
This version is built with gcc 10.2 so only runs on RISC OS hardware that has neon and vfpv3.
It also needs !SharedLibs updating to release 6. Available via PackMan.