Segmentation in image sequences : tracking human figures in motion

This thesis addressed the issue of segmentation and tracking of human figures in video sequences. Various issues affecting segmentation are investigated, including selection of a colour space representation. Different methods for segmentation are implemented and their respsective merits and disadvantages discovered. These are largely various forms of colour and motion segmentation. The different forms are then combined in order to produce a segmentation that is more effective than any of the component techniques on its own. The tracking problem is addressed separately through the use of a Smart Room, where the segmentation can be easily performed. The tracking involves a trajectory for a given person over time in the image frame co-ordinate system.

Grantor dc:publisher.institution

Department of Electrical Engineering

Year dc:date.issued

2001

Author dc:creator

• Robertson, Frances

Advisor dc:contributor.advisor

• De Jager, Gerhard

Language dc:language.iso

eng

Handle dc:identifier.uri

http://hdl.handle.net/11427/5114

OAI identifier oai:identifier

oai:open.uct.ac.za:11427/5114