A Motion Capture System Based on Natural Interaction Devices

Abstract

This thesis investigated the feasibility of building a simple and affordable motion capturing system. A pre-study was made to understand the potential uses and benefits of Natural Interaction based devices and accordingly, the Kinect sensor was used as an input device to the system. The implementation of the system was based on a client/server model. The client implemented in C++, was responsible for fetching the skeletal tracking data from the Kinect sensor, and the server, which resided in two different animation programs, implemented in Python and was responsible for making the skeletal data comprehensible to the animators who are the users of the system. The server program was extended to allow the capturing, recording, and retargeting of the motion data acquired through the sensor. The communication between the client and server was based on the Open Sound Control protocol. Furthermore, the system was tested for both functionality and usability by a sample from the target users.