Share Email Print
cover

Proceedings Paper

Deep person re-identification in aerial images
Author(s): Arne Schumann; Tobias Schuchert
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

Person re-identification is the problem of matching multiple occurrences of a person in large amounts of image or video data. In this work we propose an approach specifically tailored to re-identify people across different camera views in aerial video recordings. New challenges that arise in aerial data include unusual and more varied view angles, a moving camera and potentially large changes in environment and other in uences between recordings (i.e. between flights). Our approach addresses these new challenges. Due to their recent successes, we apply deep learning to automatically learn features for person re-identification on a number of public datasets. We evaluate these features on aerial data and propose a method to automatically select suitable pretrained features without requiring person id labels on the aerial data. We further show that tailored data augmentation methods are well suited to better cope with the larger variety in view angles. Finally, we combine our model with a metric learning approach to allow for interactive improvement of re-identification results through user feedback. We evaluate the approach on our own video dataset which contains 12 persons recorded from a UAV.

Paper Details

Date Published: 24 October 2016
PDF: 9 pages
Proc. SPIE 9995, Optics and Photonics for Counterterrorism, Crime Fighting, and Defence XII, 99950M (24 October 2016); doi: 10.1117/12.2241652
Show Author Affiliations
Arne Schumann, Fraunhofer IOSB (Germany)
Tobias Schuchert, Fraunhofer IOSB (Germany)


Published in SPIE Proceedings Vol. 9995:
Optics and Photonics for Counterterrorism, Crime Fighting, and Defence XII
Douglas Burgess; Gari Owen; Henri Bouma; Felicity Carlysle-Davies; Robert James Stokes; Yitzhak Yitzhaky, Editor(s)

Video Presentation

Deep-person-re-identification-in-aerial-images



© SPIE. Terms of Use
Back to Top