Detecting Abandoned Luggage using Bayesian Inference and Finite State Automaton(Internationa Session 7)

概要

論文の詳細を見る
This paper provides a novel approach to detect abandoned luggage by a person automatically in public places. Our proposed system relies on three modules which are moving object detection and classifying, tracking and event recognition. Especially, we focus on event recognition module in this paper. Definition of the abandoned luggage describes in event recognition module. It is composed of two sub events which are drop off luggage and unattended luggage. To detect the abandoned luggage event, we construct the Bayesian network and use finite state automaton which shows the spatio-temporal relation of two sub events. The proposed algorithm showed good results on the real world environment and also worked at real time speed.
一般社団法人電子情報通信学会の論文
2007-10-18