Bharatanatyam Hastas Recognition using CNN and Computer Vision

Bharatanatyam, one of India’s oldest classical dance forms, heavily relies on distinct hand gestures known as hastas, which play a crucial role in storytelling and expression. This research proposes a deep learning-based approach for automatic recognition of these gestures using Convolutional Neural Networks (CNN) and computer vision techniques. By training on a custom dataset of 20 unique hastas, the system learns to identify gestures with notable accuracy. The proposed model aims to support digital archiving, real-time feedback in dance education, and rehabilitation therapy involving hand movements. Experimental results demonstrate the model’s potential in recognizing complex gestures, opening up possibilities for further integration in cultural preservation and healthcare technologies.