Bharatanatyam Mudra Classification using CNN

India is the home of ancient art and one of the few nations in the world with a rich and varied cultural heritage. State by state variations exist in the nation’s languages, dances, and music.In the current technological era, there are various methods for learning these dances. The majority of people pick up the dance by going to classes. However, this method of learning is not accessible to those who are Deaf/Dumb or who choose to learn on their own. This research uses convolutional neural networks (CNNs) to provide a unique way of classifying Bharatanatyam mudras, which are complex hand movements that are essential to the Indian traditional dance style Bharatanatyam.The trained model using EfficientNetB2 architecture has shown strong performance with a test accuracy of 94 percent, indicating its effectiveness in classifying images compared to Resnet50. This study highlights the applicability of contemporary technology in maintaining and comprehending cultural history, contributing to the nexus between deep learning and traditional art forms.