This project aims to explore the language of different bodies in the field of dance by analyzing the “habitual patterns” of dancers from different backgrounds or vernaculars. For the sake of this research, the term habitual patterns is defined as the maximum likelihood of a pattern or posture that the dancer performs in improvisational dance. Framed uses statistical tools to identify the dominant cluster in a given dataset and discovers the habit within this cluster.
The video below shows the custom motion capture data collected from three dancers of different vernacular roots.
The focus lies in exposing the movement vocabulary of a dancer to reveal his/her unique fingerprint.
The proposed approach for uncovering these movement patterns is to use a clustering technique; mainly k-means. In addition to a static method of analysis, this project uses an online method of clustering using a streaming variant of k-means that integrates into the flow of components that can be used in a real-time interactive dance performance. The computational system is trained by the dancer to discover identifying patterns and therefore it enables a feedback loop resulting in a rich exchange between dancer and machine. This can help break a dancer’s tendency to create similar postures, explore larger kinespheric space and invent movement beyond their current capabilities.