Yoga Pose Detection and Identification Using MediaPipe and OpenPose Model

Yoga pose estimation is a crucial component of human pose estimation and it aids fitness freak in honing their yoga poses and avoiding harmful postures. This study describes a method for precise yoga posture detection using OpenCV and mediapipe. In the computer vision library OpenCv’ has variety of functions for video and image processing. And for estimating human posture many pretrained models are used and here MediaPipe is also one of the excellent machine learning frameworks that offers pre-trained models for estimating human posture. The suggested remedy combines the benefits of 2 techniques for creating yoga posture estimation system. The first input to the system takes a user's video and next it processes in MediaPipe to identify landmarks in the human body. To assess the user posture OnenCV is used to compute the angles between the identified landmarks. The technology offers the user immediate feedback on their posture and proposes corrective actions. The repetitions for bicep curls served as the use case for this study. The suggested system can be evaluated using a variety of challenging poses, including malasana, uktasana and jalasana. In various lighting situations, it can effectively estimate the user's position and is resistant to obstructions and background clutter. Fitness enthusiasts can use the method to identify their posture and improve their technique and form thereby lowering their chance of injury.