Yoga Pose Corrector Using Deep Learning Techniques
Dileep Kumar Kadali, K. Akhila, Mohammad Sheema, K. Gayathri, M. Charmini Ratnam, K. Satya Priyank · 2025
The increasing popularity of yoga to improve mental and physical well-being has led to a growing need for accurate and accessible guidance on performing yoga poses correctly. Incorrect execution of yoga poses can lead to ineffective practice or even injury. This paper aims to develop a Yoga Pose Corrector, a deep learning-based system designed to assist users in performing yoga poses accurately and safely. The system leverages computer vision techniques and deep learning algorithms to analyze the user’s body posture in real time. Using a pre-trained Convolutional Neural Network (CNN), the system identifies and classifies various yoga poses. It then compares the user’s pose with an ideal pose from a dataset of correctly performed yoga positions. It is the bridge to the gap fill between traditional yoga practice and modern technology. Build a platform for individuals to get guided through their positions for yoga, offering an innovative solution for yoga practitioners of all levels to improve their practice, prevent injuries, and achieve better outcomes. Implementing the Yoga Pose Corrector promises to make yoga more accurate, effective, and enjoyable.