i know some people here on dc do quite a bit of OCR, and i think it may be that even this specific question got asked in the past. would be interesting to hear what they say. my screenshot taking program (Screenshot Captor) has support for scanning, and it can be told to only grab a particular region of the screen or scanner bed, and run external tools after capturing, so that might come close, but it does seem like this would be something that calls for a dedicated program.
do the dedicated OCR applications not have something like this? what's missing from the existing applications?