Comparative Analysis of Methods of Yoga Pose Classification

Yoga practice offers numerous health benefits, and the accurate monitoring of yoga poses is pivotal for effective guidance and performance evaluation. This paper introduces a novel deep learning model for real-time prediction of yoga poses from video feeds. Leveraging state-of-the-art techniques in pose estimation, our model demonstrates robustness across diverse datasets, capturing the nuances of varied poses in different environments. We present a comprehensive methodology encompassing dataset collection, preprocessing, and model training, highlighting the model’s adaptability to real-world scenarios. Experimental results showcase the effectiveness of our approach, achieving competitive accuracy metrics. Furthermore, we seamlessly integrate the model into a real-time system, allowing for instant feedback and visualization of predicted yoga poses. This work contributes to the intersection of computer vision and fitness monitoring, fostering advancements in interactive and personalized yoga training. The presented model not only enhances the accessibility of yoga practice but also sets the stage for broader applications in gesture recognition and human-computer interaction.