Hand Gesture Detection Using Deep Learning in Bharat Natyam

Bharat Natyam is an old Indian classical dance that dates back many years. This distinctive traditional dance has always been passed by a teacher, primarily using the rote learning method resulting in repetitive steps leaving significantly less scope for learning new actions or gestures with precision. We have developed a model to represent BN hand gestures through 2 unique hand mudras. Dancer’s motions can be rebuilt by taking the features from their photos. This paper’s proposed convolution neural network is more efficient than the elective at identifying dance movements and learning on the CNN and Siamese model using different layers and predicting the accuracy of the train, validation, and test sets using other regularization techniques.