What Goes Where: Predicting Object Distributions from Above

Connor Greenwell, Scott Workman, Nathan Jacobs

In this work, we propose a cross-view learning approach, in which images captured from a ground-level view are used as weakly supervised annotations for interpreting overhead image…