CaltechAUTHORS
  A Caltech Library Service

Crosstalk Cascades for Frame-rate Pedestrian Detection

Dollár, Piotr and Appel, Ron and Kienzle, Wolf (2012) Crosstalk Cascades for Frame-rate Pedestrian Detection. In: Computer Vision – ECCV 2012. Vol.2. Springer , Berlin, pp. 645-659. ISBN 978-3-642-33708-6. https://resolver.caltech.edu/CaltechAUTHORS:20150828-112710613

[img] PDF - Accepted Version
See Usage Policy.

1472Kb

Use this Persistent URL to link to this item: https://resolver.caltech.edu/CaltechAUTHORS:20150828-112710613

Abstract

Cascades help make sliding window object detection fast, nevertheless, computational demands remain prohibitive for numerous applications. Currently, evaluation of adjacent windows proceeds independently; this is suboptimal as detector responses at nearby locations and scales are correlated. We propose to exploit these correlations by tightly coupling detector evaluation of nearby windows. We introduce two opposing mechanisms: detector excitation of promising neighbors and inhibition of inferior neighbors. By enabling neighboring detectors to communicate, crosstalk cascades achieve major gains (4-30x speedup) over cascades evaluated independently at each image location. Combined with recent advances in fast multi-scale feature computation, for which we provide an optimized implementation, our approach runs at 35-65 fps on 640 x 480 images while attaining state-of-the-art accuracy.


Item Type:Book Section
Related URLs:
URLURL TypeDescription
http://dx.doi.org/10.1007/978-3-642-33709-3_46DOIArticle
http://link.springer.com/chapter/10.1007%2F978-3-642-33709-3_46PublisherArticle
Additional Information:©2012 Springer.
Record Number:CaltechAUTHORS:20150828-112710613
Persistent URL:https://resolver.caltech.edu/CaltechAUTHORS:20150828-112710613
Usage Policy:No commercial reproduction, distribution, display or performance rights in this work are provided.
ID Code:59956
Collection:CaltechAUTHORS
Deposited By: Caroline Murphy
Deposited On:28 Aug 2015 22:51
Last Modified:03 Oct 2019 08:51

Repository Staff Only: item control page