Yoga Trainer with AI

Computer Science And Engineering

Yoga is beneficial for everyone of any age's physical and mental well-being. To avoid damaging the bones, muscles, and ligaments, it is extremely important to execute yoga poses correctly when practicing alone. As a result, providing feedback to the performance without a live instructor is possible by using artificial intelligence and machine learning in conjunction with picture processing. In addition to correcting users who perform the yoga posture improperly, the suggested system is designed to advise users on how to do it correctly on real-time and there is also a pose classification feature that can classify the yoga poses based on the images of the final yoga pose which is provided by the user. The feedback, which comes in text and audio formats, can assist the practitioner avoid injuries while also maximizing the advantages of doing the yoga posture. Various images from the internet were used to create the data set of various yoga poses. The data points are extracted from each webcam image with the aid of OpenCV and media pipelines. This is now fed into a deep learning model that employs convolutional neural networks (CNN), which finds pose errors and calculates the error percentage before providing the user with the necessary feedback in the form of text or audio desired output with a classification accuracy of roughly 95%.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.