Yoga Pose Detection Using Mediapipe and Cue Method

Popular exercise like yoga has numerous positive health effects. However, incorrectly executed yoga poses can be harmful and lessen the advantages of the practice. In recent years, yoga positions have been investigated and altered using computer vision techniques. This study introduces a novel Mediapipe-based cue-based computer vision method for improving yoga postures. By finding and recognizing the most crucial body aspects in yoga postures using deep learning algorithms, the suggested method is contrasted with proper alignment. The user is then given instructions on how to change the stance to attain the desired alignment right away by the computer. The proposed method is tested using the dataset of yoga poses, and the accuracy and real-time performance outcomes are promising. In yoga classes, online tutorials, and other situations, the suggested technique might be employed to maximize benefits while lowering risks of injury for practitioners.