Flávio Sales Truzzi,  Nicolau Leal Werneck,  Anna Helena Reali Costa. 

This paper presents a system for performing visual tracking of people running or walking in image sequences captured by static cameras. The system is divided into two parts: detection, where an adaptive model of background scene is subtracted from the current image, and tracking using Kalman filter, where people are modeled in a simple way to ensure efficiency. Tests were performed in the tracking of soccer players and people monitored by security cameras, showing the effectiveness of the proposal.

http://www.lbd.dcc.ufmg.br/colecoes/wvc/2010/0043.pdf

Processando consulta.