Still OCR: Recognizing large characters

[Shun]

Hello Mike,

when I pass an enlarged screenshot containing Chinese characters to the Pleco 4.0 beta 31 Still OCR module, the OCR recognizer splits the characters into subcharacters.

The full image:



Recognizing the top half of 皇帝:



Recognizing the bottom half:



It doesn't happen with normal-sized characters.

Could there be a way to make Pleco recognize large characters?

Cheers, Shun

 

[mikelove]

Does it work better if you change the algorithm setting (gear icon) to one of the other options?

 

[Shun]

Yes: The choices "pleco" and "apple" work fine, only the default "combined" algorithm seems to have trouble.