Panel: Motion Capture and Dance: what it can do, what it can’t do, and what it should never attempt
Movement analysis in the scientific realm is often characterised by the desire for certainty and predictability yet this level of predictability may not always be desirable in an artistic context. In the area of dance creation and performance, the fact that humans can be surprising, inspired, fragile and unpredictable could be seen as part of their potential creativity and preferable to obvious, robotic responses. How far would we allow machines to carry similar traits in their analysis and responses to human movement? This presentation investigates the area of machine learning and understanding as it may be applied to motion capture analysis in performance, including the aspiration of many machine learning techniques to model and approach human learning.
- John McCormick (AU) has been active in the area of dance and new media for many years. He was a founding member of Company In Space (with Hellen Sky), Dancehouse and Squaretangle (with Adam Nash). He has worked on a number of performances incorporating motion capture technologies including the Company In Space works CO3 (2001), The Lightroom (2004) and Sentient Space (2005), the latter in collaboration with igloo and Adam Nash, Aura (2009) with Kim Vincs and Choreotopography (2010) with Kim Vincs, Daniel Skovli, Peter Divers and Rob Vincs. John teaches motion capture and dance at Deakin University in Melbourne, Australia and is currently undertaking research at the Deakin Motion.lab and the Centre for Intelligent Systems Research.
Full text (PDF) p. 1667-1669 [Different title!]