Recognition of Indian Classical Dance Hand Gestures
R. Pradeep, R Rajeshwari, V R Ruchita, Radhika Bubna, H. R. Mamatha
Abstract
The use of hand gestures to convey events or objects visually is one of the most distinctive features of Indian classical dance. These gestures can be performed either single handedly or with both hands. This work investigates the potential of recognizing hand gestures, or mudras, of the various classical dance forms of India. By applying a variety of pre-trained CNN models to train the dataset, it seeks to test the efficiency of convolutional neural networks at distinguishing and categorizing hand gestures from images. It also involves evaluating the accuracy of each model and analyzing the performance metrics of the same. The article also entails the lesser explored domain of recognizing mudras from dance videos using object detection techniques, and combining it with the results obtained from the CNN model.