People re-identification using single or multiple camera acquisitions constitutes a major challenge in visual surveillance analysis. The main application of this research field consists to reacquire a person of interest in different non-overlapping locations over different camera views. This paper present an original solution to this problem based on a graph description of each person. In particular, a recently proposed graph kernel is used to apply Principal Component Analysis (PCA) to the graph domain. The method has been experimentally tested on two video sequences from the PETS2009 database.
People re-identification by Graph Kernels Methods
CONTE, Donatello;FOGGIA, PASQUALE;VENTO, Mario
2011-01-01
Abstract
People re-identification using single or multiple camera acquisitions constitutes a major challenge in visual surveillance analysis. The main application of this research field consists to reacquire a person of interest in different non-overlapping locations over different camera views. This paper present an original solution to this problem based on a graph description of each person. In particular, a recently proposed graph kernel is used to apply Principal Component Analysis (PCA) to the graph domain. The method has been experimentally tested on two video sequences from the PETS2009 database.File in questo prodotto:
Non ci sono file associati a questo prodotto.
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
