UPDATE 9/11/17: It does a fair OCR job, for both languages (Greek & english) Only in some cases. Not always. It depends from the text on the picture I guess. The translation fails though, But I really don't need the translation and I have disabled it from the Options. I just need the OCR alone.
This tool is Not good for capturing URL links. Because I found this screenshot which is an example of the "error" popup window. from one web page and the text was: [[<>]]
I tried to OCR this, But it could only capture only the simple 2 lines of text: [The Requested file was not a valid file]
BUT the lines which have special characters couldn't recognize it: [C:\Users\.....]

* I have also tested it on the "dilbert.com" website (like you did on your example) and it works capturing text on Comic images.
I have noticed that sometimes It converts the (i) letter as a Number (1) and the (comma) as a (full stop dot) in some images.

* You should add the (Options Button) when we right Click on the Addon. It is a painful process to find the options to change the Language.

* Also on the Youtube video in (FIREFOX 56.0.2) I tried your Chinese example video and it works Great! => https://youtu.be/8dRWw4FIYqs
But on YouTube video, in the (CHROME Browser) it is Unable to capture the text. it shows me this Warning:
(Please try again when the page has loaded. Prevent this page from creating additional dialogs.)

*And last but not least, On another Comic Image Example it can not recognize this kind of text/font => http://www.leasticoulddo.com/comic/20170430/

Anyway, all in all, this Addon works fairly with Firebox 56.0.2!
I want to believe that in Future updates will be much better. I guess.