Posture Recognition in Bharathanatyam Images using 2D-CNN

Abstract

The postures are important for conveying emotions, expressing artistic intent, and preserving appropriate technique. Posture recognition in dance is essential for several reasons, as it improving the performance and overall artistic expression of the dancer. The Samapadam, Aramandi, and Muzhumandi are three postures that serve as the foundation for the Bharathanatyam dance style. This work proposes a model designed to recognize the posture portrayed by the dancer. The proposed methodology employs the pre-trained 2D-CNN model fine-tuned using the Bharathanatyam dance image dataset and evaluates the model performance.