An automated stick figure generation for BharataNatyam dance visualization

Abstract

BharataNatyam (BN) is an ancient Indian Classical Dance dating centuries ago. This unique classical dance has been taught by a teacher to a student mostly by rote learning method. A student uses various methods to record the dance choreography taught by a teacher. Currently recording through mobile phones and various other devices of a live dance performance are popular methods but it has its own inherent disadvantages. Several attempts of automation in choreography are reported and dance visualization is the key factor. Stick Figure representation is a popular method amongst all and is still being used by many of the practitioners. We have developed a model to represent a BN dance step through a unique thirty attribute dance position vector. Evolutionary approach of Genetic Algorithms generates non-conventional BN dance poses which are approved by renowned dance experts. However, in order to visualize every resulting BN dance, a human model has to pose accordingly. We overcame this hurdle by developing a stick figure generation module. In this paper we present the details of stick figure and in particular how we mold it to suit to a BN dance pose that corresponds to a Dance position vector.