CaltechAUTHORS
  A Caltech Library Service

The Fastest Pedestrian Detector in the West

Dollár, Piotr and Belongie, Serge and Perona, Pietro (2010) The Fastest Pedestrian Detector in the West. In: Electronic Proceedings of the British Machine Vision Conference BMVC 2010 ; Aberystwyth, 31st August - 3rd September 2010. BMVA Press , Durham, UK, Art. No. 68. ISBN 1901725405. http://resolver.caltech.edu/CaltechAUTHORS:20190328-144424276

[img] PDF - Published Version
See Usage Policy.

1432Kb

Use this Persistent URL to link to this item: http://resolver.caltech.edu/CaltechAUTHORS:20190328-144424276

Abstract

We demonstrate a multiscale pedestrian detector operating in near real time (~6 fps on 640x480 images) with state-of-the-art detection performance. The computational bottleneck of many modern detectors is the construction of an image pyramid, typically sampled at 8-16 scales per octave, and associated feature computations at each scale. We propose a technique to avoid constructing such a finely sampled image pyramid without sacrificing performance: our key insight is that for a broad family of features, including gradient histograms, the feature responses computed at a single scale can be used to approximate feature responses at nearby scales. The approximation is accurate within an entire scale octave. This allows us to decouple the sampling of the image pyramid from the sampling of detection scales. Overall, our approximation yields a speedup of 10-100 times over competing methods with only a minor loss in detection accuracy of about 1-2% on the Caltech Pedestrian dataset across a wide range of evaluation settings. The results are confirmed on three additional datasets (INRIA, ETH, and TUD-Brussels) where our method always scores within a few percent of the state-of-the-art while being 1-2 orders of magnitude faster. The approach is general and should be widely applicable.


Item Type:Book Section
Related URLs:
URLURL TypeDescription
https://doi.org/10.5244/c.24.68DOIArticle
ORCID:
AuthorORCID
Belongie, Serge0000-0002-0388-5217
Perona, Pietro0000-0002-7583-5809
Additional Information:© 2010. The copyright of this document resides with its authors. It may be distributed unchanged freely in print or electronic forms. P. P. was supported by ONR MURI Grant #N00014-06-1-0734. S. B. was supported in part by NSF CAREER Grant #0448615 and ONR MURI Grant #N00014-08-1-0638.
Funders:
Funding AgencyGrant Number
Office of Naval Research (ONR)N00014-06-1-0734
NSFIIS-0448615
Office of Naval Research (ONR)N00014-08-1-0638
Record Number:CaltechAUTHORS:20190328-144424276
Persistent URL:http://resolver.caltech.edu/CaltechAUTHORS:20190328-144424276
Official Citation:Piotr Dollar, Serge Belongie, and Pietro Perona. The Fastest Pedestrian Detector in the West. In Frédéric Labrosse, Reyer Zwiggelaar, Yonghuai Liu, and Bernie Tiddeman, editors, Proceedings of the British Machine Vision Conference, pages 68.1-68.11. BMVA Press, September 2010. doi:10.5244/C.24.68.
Usage Policy:No commercial reproduction, distribution, display or performance rights in this work are provided.
ID Code:94253
Collection:CaltechAUTHORS
Deposited By: George Porter
Deposited On:28 Mar 2019 22:44
Last Modified:28 Mar 2019 22:44

Repository Staff Only: item control page