Not able to recognize single character

[David Roberts]

Hi all,

Basically, in conducting image reformatting to get rid of noise around digits, I've disjoined the top my 3 char, causing Tesseract to return empty strings. Does anyone have any advice or time-efficient solution for getting tesseract to recognize the 3 digit? 

Currently, I'm using -psm 7 with the "digits" argument for configuration.

I hesitate to change my preprocessing because the images I'm pulling from have a lot going on in the background and are very diverse in terms of shading and coloring. My current method works... mostly, so I was hoping to make changes on the Tesseract side.

Best,

David