There isn't a mechanism for telling the tracker to only search one portion of the camera image for trackables. I wonder, would you prefer a method for rendering the camera image to only one part of the screen? Then you wouldn't be hiding portions of the camera image with your UI, instead you would shrink the camera image to fit a smaller portion of the screen.

- Kim

A method to render only a portion of the 3D camera output would be helpful, since rendering an empty area that is obscured by the UI is a bit of a waste. That would be great, thanks.

This doesn't address the initial question regarding defining/limiting the screen area for tracking.

Hypothetically, if the majority of the data being passed to the tracker was black with only a swath in the center containing image data, would this speed up the tracker's ability to acquire and analyze the incoming camera data?

Thanks,
-liam m-

I expect that it would, the question is whether this efficiency is overcome by the cost of preprocessing the frame.

Rather than isolating the image in a black field, could you expand it to fill the entire frame? So only the target image is presented to the tracker.

Keep in mind that the tracker is analyzing the camera image, the screen image is the result of multiple render steps and transformations.

The answer to that sounds like a fun little research project. I'd like to try this after our current project gets out the door.

I guess the ability to define the trackable area would go into the 'wish list but not essential' pile.