Classification Of Yoga Pose Using Pretrained Convolutional Neural Networks

Abstract

Yoga is an extraordinary spiritual science of self-development and self-realization that shows us how to develop our full potential in our many-sided lives. Yoga has become a way of life for many people all over the world in recent years. As a result, a scientific analysis of yoga postures is required. Recognizing posture is a difficult task due to the scarcity of datasets and the need to detect posture from a huge dataset. To address this issue, a large dataset containing five different types of yoga pose images was collected from the Kaggle online dataset. In the proposed work, the yoga pose is taken as input, the features were detected using, pre-trained models MobileNetV2 and DenseNet201. The Classifiers namely Support Vector Machine (SVM) and Random Forest (RF) are used for classification. By comparing the experimental results the performance of MobileNetvV2 with Random Forest yields better results when compared to other models.