MobileRead Forums - View Single Post

roger64 · 03-05-2016, 08:12 AM

Hi

This approach is very promising and interesting as it allows to correct some recurrent OCR mistakes (after a dictionary check).

In French, we have a list of hundreds of words that deserve to be checked out of OCR. With such an approach, of course, you can't avoid false positives, which means we need not to correct them but just to highlight them, to be able later to speed up a manual checking.

I give one example, if you find "trame" which is a correct but fairly rare word, 98% of the time, it should be written "traine". It makes sense to highlight it.

However, I do not know how to write a single entry.

Could some kind soul write a code paragraph example of this Calibre function allowing me to highlight "trame"?

03-05-2016, 08:12 AM	#17
roger64 Wizard Posts: 2,608 Karma: 3000161 Join Date: Jan 2009 Device: Kindle PW3 (wifi)	Hi This approach is very promising and interesting as it allows to correct some recurrent OCR mistakes (after a dictionary check). In French, we have a list of hundreds of words that deserve to be checked out of OCR. With such an approach, of course, you can't avoid false positives, which means we need not to correct them but just to highlight them, to be able later to speed up a manual checking. I give one example, if you find "trame" which is a correct but fairly rare word, 98% of the time, it should be written "traine". It makes sense to highlight it. However, I do not know how to write a single entry. Could some kind soul write a code paragraph example of this Calibre function allowing me to highlight "trame"?