Using colour features to classify objects and people in a video surveillance network

Visual tracking of humans has proved to be an extremely challenging task for computer vision systems. One idea towards a development of these systems is the incorporation of colour. Often colour appearance of a person can provide enough information to identify an object or person in the short-term. However, the imprecise nature of colour measurements typically encountered in image processing has limited their use. This thesis presents a system which uses colour appearances of objects and people for tracking across multiple camera views in a digital video surveillance network.

Grantor dc:publisher.institution

Department of Electrical Engineering

Year dc:date.issued

2004

Author dc:creator

• Price, Mathew

Advisors dc:contributor.advisor

• De Jager, Gerhard
• Nicolls, Fred

Language dc:language.iso

eng

Handle dc:identifier.uri

http://hdl.handle.net/11427/5121

OAI identifier oai:identifier

oai:open.uct.ac.za:11427/5121