Investigating Human-Computer Interaction with Motion-Capture Algorithms using a Microsoft Kinect

Abstract

This bachelor thesis investigates tracking of the human body in a 3D environment using known computer-visual algorithms. This is done by evaluating a variety of filter-, detection-, and trackingalgorithms, and assessing their strength and weaknesses in different environmental conditions that are found in most indoor situations. We will also discuss different combinations of algorithms that work together to optimize efficiency in a variety of conditions with limited resources. The purpose of the developed application has been to evaluate how the algorithms perform and investigate the different possibilities to track using an ordinary laptop and a Microsoft Kinect. Finally, we discuss at what level of detail the tracking could be used and investigate if this kind of system could manage emotional capturing i.e. tracking facial expressions simultaneously with body movement.