A Novel Approach for Developing Inclusive Real-Time Yoga Pose Detection for Health and Wellness Using Raspberry pi

Yoga, an ancient practice with myriad physical and mental benefits, has gained popularity across the globe. Accurate pose detection and guidance are crucial for practitioners to maximize the benefits and minimize the risk of injury. This study introduces a novel approach to real-time yoga pose identification and instruction using Raspberry Pi. Traditionally, machine learning algorithms have been utilized for pose detection; however, this study uses MediaPipe library and OpenCV for pre-trained pose estimation model from the MediaPipe library to detect and track the key landmarks of a person's body. Based on the detected landmarks, it calculates various angles between the joints and classifies the pose accordingly. The proposed system's core advantage lies in its utilization of Mediapipe, an open-source framework developed by Google, which efficiently identifies and tracks essential body landmarks over time. These landmarks, alongside TensorFlow's capabilities, are pivotal in extracting pertinent features from images and videos. OpenCV, a renowned computer vision library, further enhances the system by processing visual data. The user interface is designed to offer user convenience and efficient interaction. By providing real-time feedback through voice commands, the system assists users in achieving proper yoga poses. The Raspberry Pi, equipped with its camera module, serves as the hardware foundation. Its affordability, adaptability, and compact size make it an ideal choice for on-device processing, ensuring low latency and responsive performance. Additionally, it provides some simple stretches and recommendations for alleviating muscular stiffness and soreness brought on by inadequate yoga or other wrongdoing movements and finally provides a searching option for yoga pose.