Classification of Indian Dance Forms using Pre-Trained Model-VGG
Snigdha Biswas, Anirudh Ghildiyal, Sachin Sharma
Abstract
Indian classical dance often referred to as ICD acts as a mirror of the rich cultural heritage of India. Every dance comes from a different state in the country. The national level academy for performing arts in India called The Sangeet Natak Academy identifies eight different dance forms namely- `Bharatanatyam', `Kathak', `Kathakali', `Kuchipudi', `Manipuri', `Mohiniyattam', `Odissi' and `Sattriya'. In this paper, we have used transfer learning to solve this multiclass problem. Pre-trained models such as VGG16 and VGG19 were used to classify the images of the 8 dance forms known in India. This was completed using the ICD data set from Kaggle consisting of 8 different classes and multiple images in each class.