Combine Comic Text Detector with Google Lens?

[Keksus1]

Would it be possible to somehow combine the Comic Text Detector with Google Lens? The idea here would be that only the part of the image detected by Comic Text Detector would be sent to Google Lens which would reduce the amount of data one would need to send. Something people with slower internet connections would benefit from. At least in my case the experience is that sending a full 1440p image to Google Lens for analyzation can ... take a while.

In general with templates it doesn't seem to big of a problem when the text area is always in the same place. But in some games text can pop up anywhere which reduces the efficiency of AutoOCR unless I use the entire image of a game as the OCR area.