International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 12 Issue: 05 | May 2025
p-ISSN: 2395-0072
www.irjet.net
Yoga Steps Prediction and Correction using Computer Vision Sweta Kale1, Rakesh Rajput2, Aarti Khandagale3, Payal Musale4, Akshay Thorat5 1Professor, Dept. of Information Technology, RMD Sinhgad School of Engineering, Warje, Pune, Maharashtra, India 2,3,4,5Student, Dept. of Information Technology, RMD Sinhgad School of Engineering, Warje, Pune, Maharashtra,
India -------------------------------------------------------------------------***-----------------------------------------------------------------------ABSTRACT Yoga is a well-liked activity that incorporates breathing techniques and a variety of bodily positions to enhance both mental and physical well-being. In this project, we have created a system that uses a webcam and machine learning algorithms to identify yoga positions in real time. Bhadrasan, Shavasan, Gomukhasan, Vajrasan, Sarvangasan, Shirsansan, Chakrasan, and Dhanurasan are among the eight yoga positions used in the method. The first step in the system’s operation is gathering a dataset of yoga positions together with the virtual coordinates that correlate to each stance. Using 501 virtual coordinates on the yoga practitioner’s face and body, we manually marked the poses. Normalizing the coordinates and dividing the dataset into training and testing sets were preprocessing steps.For each pose, we extracted features from the virtual coordinates using the MediaPipe and OpenCV libraries. We utilized the OpenCV and MediaPipe libraries to record a live webcam stream and extract virtual coordinates from the video frames after the model had been trained and verified. The trained model was then used to predict the matching yoga stance based on the retrieved coordinates. The name and accuracy of the predicted pose are shown on the screen in real time by the system. We used a SQLite database and the tkinter package to create a user registration and login system in addition to pose detection. Users were able to register their information through this method, and it was saved in the database for use in subsequent logins. Yoga practitioners can use the system to track and enhance their practice, and teachers can use it to keep an eye on their pupils and provide them tailored feedback. To increase the system's accuracy, further work will involve adding feedback mechanisms and enlarging the dataset to include more yoga postures.
Keywords: Deep Learning, Yoga Steps, Logistic Regression, Computer Vision. 1. INTRODUCTION Yoga is a popular practice that involves various body postures and breathing exercises to improve physical and mental health. The correct execution of yoga postures is essential to reap the full benefits of the practice. However, it can be challenging to identify the correct alignment and posture while practicing yoga, especially for beginners. Therefore, there is a need for a system that can detect, and monitor yoga poses in real-time. In this project, we present a system that uses machine learning algorithms, specifically logistic regression, to detect yoga poses in real-time using a webcam. Our system is capable of detecting eight different yoga poses, including Bhadrasan, Shavasan, Gomukhasan, Vajrasan, Sarvangasan, Shirsansan, Chakrasan, and Dhanurasan. Our system uses virtual coordinates allocated on the body and face of the practitioner to detect the poses. We collected a dataset of yoga poses with corresponding virtual coordinates for each pose. The dataset was preprocessed, and logistic regression model was trained to classify the poses based on the extracted features. Once the model was trained and validated, we used OpenCV and MediaPipe libraries to capture live video feed from the webcam and extract virtual coordinates from the video frames. The extracted coordinates were then passed through the trained logistic regression model to predict the corresponding yoga pose. The system displays the name and accuracy of the predicted pose on the screen in real-time. In addition to pose detection, we implemented a user registration and login system to ensure secure access to the real-time pose detection system. The proposed system can be used by yoga practitioners to monitor and improve their practice, as well as by instructors to monitor their students and provide personalized feedback. 1.1 MOTIVATION
Self-Learning Assistant System needs to support the self-learners in receiving feedback for correct experiences and exercises.
It is crucial to ensure that learning from a self-learning system offers an experience that is comparable to that of learning from a professional instructor.
© 2025, IRJET
|
Impact Factor value: 8.315
|
ISO 9001:2008 Certified Journal
|
Page 356