Publication

Abstract : Yoga has proved to be the oldest and the most effective practice to attain a healthy life. Over the years its popularity is only increasing day by day. A lot of young people feel fascinated and find it an effortless way to per sue a good mental health. It helps in energizing and calming of the one who is constantly practicing it. That is why there is a need for a yoga position detection system. We tend to come with a new solution because the already available ones only instruct how the poses can be done it does not have the ability to monitor the user while doing. So, it forms a drawback that the user cannot be assisted or been notified weather he is practicing for the right manner that it is supposed to be performed. So, our proposed system is used to detect the position infer the accuracy of the positions practiced to significantly lead the user to outreach the right posture while one is frying to extract the best outcomes from doing the asanas. The proposed model uses Google's pose net model to categorize the poses by forming its own skeleton image and then comparing it with the pre trained dataset consisting of various yoga asanas. This will keep up the time, continuous monitoring, and accuracy measure to find and drive the user to follow the right positions. This model will help the people extract the maximum outcome without the need for any trainer or external assistance. We tend to come with a new solution because the already available ones only instruct how the poses can be done it does not have the ability to monitor the user while doing.