Yoga Pose Identifier and Gamification using MediaPipe

Welcome to the Yoga Pose Identifier! This project uses MediaPipe to detect and classify yoga poses in real-time. The system not only recognizes the pose but also gamifies the experience by providing real-time feedback and scoring based on the accuracy of the pose. Also providing me big help was this kaggle dataset, that had the yoga poses: https://www.kaggle.com/datasets/niharika41298/yoga-poses-dataset

Overview

This project leverages the power of MediaPipe's pose estimation technology to track and evaluate yoga poses. The system can:

1. Identify Yoga Poses: Detects common yoga poses such as the "Tree" and "Goddess" poses by comparing live camera input or images against pre-trained reference poses.
2. Provide Real-Time Feedback: Offers real-time feedback on the accuracy of the pose, with a score and streak system to gamify the experience.
3. Custom Pose Training: Allows for training on custom poses using reference .npy files that store the landmarks of correctly performed poses.

Features

• Pose Classification: Compares the user's pose to reference poses (e.g., "Tree", "Goddess") and provides immediate feedback.
• Gamification: Adds a score and streak mechanism to encourage correct pose maintenance and improvement.
• Visualization: Displays the pose landmarks on the user's image, helping them visualize how well they are matching the reference pose.