Chapter

Human Behavior Understanding in Networked Sensing

pp 283-298

Date:

Exploiting Crowd Synthesis for Multi-camera Human Tracking

  • Zhixing JinAffiliated withThe Center for Research in Intelligent Systems, University of California Email author 
  • , Bir BhanuAffiliated withThe Center for Research in Intelligent Systems, University of California

* Final gross prices may vary according to local VAT.

Get Access

Abstract

There are many challenges to achieve a robust performance for tracking in a video network. In this chapter, we propose a method that integrates both detection and crowd synthesis approaches to achieve robust tracking performance. The experiments are conducted on PETS 2009 data set, and the performance is evaluated by multiple object tracking precision and accuracy criteria based on the position of each pedestrian on the ground plane. It is demonstrated that the information from crowd synthesis can provide significant advantage for tracking multiple pedestrians through multiple cameras.