Home // UBICOMM 2015, The Ninth International Conference on Mobile Ubiquitous Computing, Systems, Services and Technologies // View article

A Robust Model for Person Re-Identification in Multimodal Person Localization

Authors:
Thi Thanh Thuy Pham
Thi Lan Le
Trung Kien Dao
Duy Hung Le
Van Toi Nguyen

Keywords: Multimodal person localization; Person Re-ID; KDES descriptor; Camera; WiFi.

Abstract:
Determining person ID (Identity) is one of the crucial steps in indoor human localization system. It is more exactly stated as person Re-ID (Re-Identification) problem because for each user’s position, the user ID at the first occurrence needs to be shown correspondingly at the later times of localization. In this paper, a multimodal person localization system of WiFi and camera is proposed, with the analysis for the key role of appearance-based person Re-ID in fusing different information sources. A new model for person appearance representation based on kernel descriptor is proposed to tackle the challenges of person Re-ID in camera network. Additionally, a dataset for the real scenario of the proposed multimodal person localization is also established, in which we set a database for vision-based human Re-ID evaluation. The experiments on the benchmark datasets and our dataset show the outperforming results in comparison with other state-of-the-art approaches.

Pages: 38 to 43

Copyright: Copyright (c) IARIA, 2015

Publication date: July 19, 2015

Published in: conference

ISSN: 2308-4278

ISBN: 978-1-61208-418-3

Location: Nice, France

Dates: from July 19, 2015 to July 24, 2015