Adriano Gimenez Fernandez Undergraduate Dissertation 2015/16

Using Human Motion Capture to Control the Movement of an Elephant

Supervised by S.Maddock

Abstract

Animation is a process that has been simplified greatly by the use of computers. From artists rendering every image in a scene, to computers using motion capture to read the movements of an actor and transfer them onto a three-dimensional model on a computer, the aim is to provide an entertaining platform where people can enjoy this medium through movies and video games. Animals such as elephants that have a complex structure called muscular hydrostat (their trunk) are a lot more complex to animate than a humanoid character. This is because muscular hydrostats can elongate, twist and curve whilst being entirely composed of muscle fibres .

The aim of the project is to create a system that can successfully animate an elephant with the use of Microsoft's Kinect . The Kinect is an economically friendly and easy to use camera that is easily accessible to a large audience. The camera captures and tracks your body movement and maps it to a model elephant on the computer. There are three different techniques that will be investigated. The first method will map each different part of both arms to different sections of the trunk. The second method will solely map one arm to the trunk of the elephant. The last method will look at the elongation of the trunk and a way to control it in conjunction with the second technique. The elongation is controlled by the height of the left hand from the floor. If it goes above 1.2 metres the trunk will elongate and if it falls below 0.9 metres , it will shrink.

From the survey, it was clear that the participants were engaged with the program and were very intrigued with the different mapping techniques that were presented. There was, however, a learning aspect that they all seemed to have. All of the participants preferred the last mapping technique they tried. This indicates that due to the time that they spent on the kinect , the end of the experiment, they became more comfortable with the range of motion available. The results indicated that the best mapping technique was the two arms mapping technique. This is due to the wide range of motion available and how simple it was to interact with it.