How are bounding boxes interpreted?

I followed the basic docs to create the FaceDetector. However, I'm not quite sure how the bounding box values are evoked, because the values I get are out of range for every mobile device. 

Example: 

`Rect.fromLTRB(1285.0, 2859.0, 3054.0, 4627.0)`

I want to add the bounding box to the taken image. 

Am I missing something? 
