Object Detection Method Based on Local Kernels and Automatic Kernel Selection by Kullback-Leibler Divergence

Authors:
Kazuhiro HOTTA
Affiliations:
-
Venue:
WACV '02 Proceedings of the Sixth IEEE Workshop on Applications of Computer Vision
Year:
2002

Citing 0
Cited 1

A novel and stable human detection and behavior recognition method based on depth sensor

3D Research

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper presents a object detection method basedon local kernels. The local kernels are arranged to allpositions on recognition target and are selected automaticallyby using Kullback-Leibler divergence accordingto the recognition target. The proposed methodis applied to pedestrian detection problem. The performanceof the proposed method is evaluated by theexperiment using MIT CBCL pedestrian database. Itis confirmed that generalization ability of the proposedmethod is improved by selecting the local kernels automatically.