What Goes Where: Predicting Object Distributions from Above

TL;DR

This paper introduces a cross-view learning method that uses ground-level images as weak supervision to train a neural network for predicting object types and counts from overhead imagery, enabling semantic understanding without manual labels.

Contribution

The novel approach leverages weakly supervised learning from ground-level images to interpret overhead imagery, reducing the need for manual annotations.

Findings

Network captures semantically meaningful features

Effective in predicting object types and counts

Works well on large geotagged datasets

Abstract

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 imagery. The outcome is a convolutional neural network for overhead imagery that is capable of predicting the type and count of objects that are likely to be seen from a ground-level perspective. We demonstrate our approach on a large dataset of geotagged ground-level and overhead imagery and find that our network captures semantically meaningful features, despite being trained without manual annotations.