Multi-spectral Pedestrian Detection via Image Fusion and Deep Neural Networks

French, Geoffrey, Finlayson, Graham and Mackiewicz, Michal (2018) Multi-spectral Pedestrian Detection via Image Fusion and Deep Neural Networks. Journal of Imaging Science and Technology. ISSN 1062-3701

[img] PDF (Accepted manuscript) - Submitted Version
Restricted to Repository staff only until 31 December 2099.

Download (3200kB) | Request a copy

    Abstract

    The use of multi-spectral imaging has been found to improve the accuracy of deep neural network-based pedestrian detection systems, particularly in challenging night time conditions in which pedestrians are more clearly visible in thermal long-wave infrared bands than in plain RGB. In this article, the authors use the Spectral Edge image fusion method to fuse visible RGB and IR imagery, prior to processing using a neural network-based pedestrian detection system. The use of image fusion permits the use of a standard RGB object detection network without requiring the architectural modifications that are required to handle multi-spectral input. We contrast the performance of networks trained using fused images to those that use plain RGB images and networks that use a multi-spectral input. © 2018 Society for Imaging Science and Technology.

    Item Type: Article
    Faculty \ School: Faculty of Science > School of Computing Sciences
    Depositing User: LivePure Connector
    Date Deposited: 20 Sep 2018 17:30
    Last Modified: 18 Feb 2019 10:30
    URI: https://ueaeprints.uea.ac.uk/id/eprint/68324
    DOI: 10.2352/J.lmagingSci.Technol.2018.62.5.050406

    Actions (login required)

    View Item