Bug report: Optical Character Recognition (OCR) not working

[mattnotmax]

Describe the bug Current version 9.37.3 the OCR operation stalls at the 'finding text' stage. This is for Firefox and Chrome tested against the GCHQ github instance.

However, if I run a Docker version of 9.37.3 I can get OCR to work, but only on Firefox.

I have downgraded to 9.32.0 on another webserver (as I saw that tesseract was updated from 2.1.1 -> 2.1.5 in 9.32.1). I can get OCR to work but again only in Firefox and Chrome has the same issue.

To Reproduce Steps to reproduce the behaviour or a link to the recipe / input used to cause the bug:

1. Load a PNG image with text
2. Apply OCR operation
3. OCR generation starts and stalls at 'finding text'
4. OCRed text does not appear

Expected behaviour The text should be OCRed.

Screenshots

Chrome - current version - GCHQ github

Firefox 9.32.0 working

Docker current version - working

Desktop (if relevant, please complete the following information): Tested Chrome Version 100.0.4896.127 (Official Build) (64-bit) and Chrome Version 101.0.4951.41 (Official Build) (64-bit) with same results.

Tested Firefox 99.0.1 (64-bit)

Additional context Thanks as always for a great tool.

[madeinoz67]

yes confirming same issue with Edge