A Real-Time Virtual Yoga Assistant Using Machine Learning

Yoga is a practice that aims to develop an all-around personality by synchronizing the mind, body, and spirit. However, incorrect postures or techniques can cause damage. In ancient times, yoga was performed under the supervision of a teacher, but it is difficult to find a competent guru in today's fast-paced world. The goal of this project is to develop an application that can track and evaluate physical exercise, specifically yoga, through the use of human pose estimation. This application, called “A Real-Time Virtual Yoga Assistant,” uses machine learning methodologies to classify data on yoga positions in both pre-recorded and real-time videos. The research also examines various pose estimation and key point detection approaches and deep learning models used for posture classification.