Towards automation and classification of BharataNatyam dance sequences

Abstract

BharataNatyam (BN) is an ancient Indian Classical Dance (ICD). Creativity and innovation are the soul of any art including BN dance. Within the framework of rules and traditionally accepted boundaries, choreographers try to innovate and create unseen, aesthetic and novel BN dance sequences. The human e orts can be supported with the computational assistance to generate valid, genuine BN dance sequences. Moreover, these movements can be empowered by the unspecified rules extracted from the analysis of "Adavus", which are considered ideal BN dance sequences for "Nritta" or pure dance movements. Thus an altogether new and interesting sequence can be obtained. In this paper we present our experimental ArtToSMart (System Modeled art) system, which gradually enhanced from one beat dance pose generation to 'm' beat dance sequences generation comprising of 'm' system generated dance poses. Furthermore, we are tagging these sequences with the help of BN dance experts and trying to develop a machine learning model to classify system generated BN sequences. The use of Rough Set tools have proved to be impressive for the same.