Undergrad Research Project - Virtual Dancer

Spring 2015

Natalia Lesjack-Randall
Daniel Siewiorek
Project description

Our objective is to create a system that can read, classify, and respond to a dancer’s movements. In order to achieve this, we will work in the Max/MSP/Jitter environment using existing libraries, and also writing our own in C. We will use the Xbox Kinect for visual input and preliminary processing, focusing on key joint location information. From there, we will use a rigorous school of dance theory by Rudalf Laban, translated to match the tools existing in computer science, to identify the features that will be used when creating our machine learning classifiers. Once the classifier has been sufficiently trained, we will create a system to translate its output into a creative visual response to be used as the final product. The anticipated final result will be that a dancer will be able to dance in front of the Kinect, and the system will respond in real time through images projected onto the back wall of the stage. These projections will be of abstract forms that use color and space to respond to and enhance the aesthetic and intent of the dancer.

Return to project list