Open AccessModelling large scale camera networks for identification and tracking: an abstract framework
Author(s) -
Mohan Lakshmi,
Me Vivek
Publication year - 2020
Publication title -
iet computer vision
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.38
H-Index - 37
eISSN - 1751-9640
pISSN - 1751-9632
DOI - 10.1049/iet-cvi.2019.0959
Subject(s) - computer science , scalability , identification (biology) , tracking (education) , distributed computing , key (lock) , artificial intelligence , state (computer science) , real time computing , scale (ratio) , machine learning , data mining , database , computer security , psychology , pedagogy , botany , biology , physics , algorithm , quantum mechanics
In this study, the authors discuss a novel approach for multi‐camera‐based unobtrusive identification and tracking of occupants in wide‐area, multi‐building scenarios. Considering the scalability issues in adopting a centralised approach to monitor wide‐area scenarios, they proposed a distributed approach to occupant identification and tracking. The key technical idea underlying their approach is to abstract a wide‐area indoor surveillance environment using a distributed state transition system (DSTS) model, which in turn is composed of independent building‐specific state transition systems, coordinating and collaborating with each other. This study presents the details of their DSTS model and examines the temporal ordering of recognition events within the DSTS for ensuring accurate state information and responses to spatio–temporal queries. They also provide an experimental evaluation of the performance of their model using precision‐recall metrics. Their conclusion is that the DSTS model serves as an efficient mechanism for tracking occupants in wide‐area, multi‐building scenarios monitored by camera networks.