A dataset of Sattriya dance: Classical dance of Assam

Abstract

Sattriya is an Indian classical dance of Assam composed of complex hand gestures, body moments, facial expressions and background music. At a first glance for common people meaning of mudras incorporated in Sattriya dance become difficult to understand. Authors have generated a dataset of Sattriya double handed gestures performed by total fifty artists from different dance academies which can be used to explore ways to build an automated system using machine learning and deep learning algorithms. There are eight double handed gestures or mudras made available through the referred dataset. Mendeley platform is used to store the dataset and samples were collected by visiting some protruding Sattriya dance schools located in Guwahati, Assam. In this article, various tables are used to illustrate variations of the mudras in terms of appearance, title, and compared to other Indian classical dances. To segment and analyze the mudras watershed segmentation is used. In this article researchers have enumerated a few very specific characteristics of Sattriya mudra that may be helpful to academics working in various disciplines of study.

Keywords

• Double handed mudra
• Segmentation
• Machine learning
• Deep learning
• ResNet-50
• Watershed algorithm