Archangel: A Hybrid UAV-based Human Detection Benchmark with Position and Pose Metadata
Authors:
Yi-Ting Shen,
Yaesop Lee,
Heesung Kwon,
Damon M. Conover,
Shuvra S. Bhattacharyya,
Nikolas Vale,
Joshua D. Gray,
G. Jeremy Leong,
Kenneth Evensen,
Frank Skirlo
Abstract:
Learning to detect objects, such as humans, in imagery captured by an unmanned aerial vehicle (UAV) usually suffers from tremendous variations caused by the UAV's position towards the objects. In addition, existing UAV-based benchmark datasets do not provide adequate dataset metadata, which is essential for precise model diagnosis and learning features invariant to those variations. In this paper,…
▽ More
Learning to detect objects, such as humans, in imagery captured by an unmanned aerial vehicle (UAV) usually suffers from tremendous variations caused by the UAV's position towards the objects. In addition, existing UAV-based benchmark datasets do not provide adequate dataset metadata, which is essential for precise model diagnosis and learning features invariant to those variations. In this paper, we introduce Archangel, the first UAV-based object detection dataset composed of real and synthetic subsets captured with similar imagining conditions and UAV position and object pose metadata. A series of experiments are carefully designed with a state-of-the-art object detector to demonstrate the benefits of leveraging the metadata during model evaluation. Moreover, several crucial insights involving both real and synthetic data during model optimization are presented. In the end, we discuss the advantages, limitations, and future directions regarding Archangel to highlight its distinct value for the broader machine learning community.
△ Less
Submitted 8 August, 2023; v1 submitted 31 August, 2022;
originally announced September 2022.
