Original article was published by Rishabh Jain on Artificial Intelligence on Medium
APPLICATION OF MACHINE LEARNING IN YOGA
Hey! Curators as you all know that in today’s world Artificial Intelligence has become the part of our lives and it is so vibrant today that almost every technology surrounding us uses the Artificial Intelligence to just not improve but also makes it attractive and impressive which fascinates the users of the technology towards them. Almost every field in this vast and big World uses this technology for their research , faster consequences and for the innovative suggestions that can provide a new way to the research which can lead to a New Product or Technology with the Integration of Artificial Intelligence.
Artificial Intelligence has been helping in various aspects of human life. The technology has touched almost every industry be it is the medical application, mobility, healthcare or retail and now AI has made its way to the fitness industry.
Generally when we talk about fitness all think about the Jogging, Running, Walking or Cardio Exercises and the application which is used to measure the activities and collects the data throughout the day.
But there is a very ancient and also the today’s trending aspect which is known as YOGA which is not of the past but of the future. As u all can realize the benefits of just some days of Yoga. So, for making this technology more efficient and accurate Artificial Intelligence as a Technology in Yoga is also giving its Contribution.
To make Yoga even more accessible for people Zenia– a start-up founded in Belarus started using AI in order to help people find accurate ways to perform the asanas.
Zenia is the first virtual yoga assistant which uses machine learning and movement recognition technology to guide people through their exercises.
The fitness applications which are already present on mobile stores are usually video guides, where one needs to follow the instructions provided on the video and won’t be monitored personally.
Now, even if you think you are doing them accurately, there is no way of telling whether the asanas performed have the right postures. And besides, if you’ve been exercising for a long time, you know that there is more to it than just moving body parts, such as monitoring the breathing. These online videos take up a long time to get adjusted to, particularly for the beginners.
A hybrid deep learning model is proposed using convolutional neural network (CNN) and long short-term memory (LSTM) for Yoga recognition on real-time videos, where CNN layer is used to extract features from key points of each frame obtained from Open Pose and is followed by LSTM to give temporal predictions. To the best of our knowledge, this is the first study using an end-to-end deep learning pipeline to detect Yoga from videos. The system achieves a test accuracy of 99.04% on single frames and 99.38% accuracy after polling of predictions on 45 frames of the videos.
- With this application, one can convert their home into an exercising studio where the app will monitor the person throughout your asanas.
2. It doesn’t just watch you, but it also guides you step by step, explaining the exact way to stand or bend for a particular position.
3. It also tells the user the exact feeling one should be feeling during the exercise with voice-controlled commands.
4. The app has introduced 40 basic and mid-level asanas as of now and moving forward it aims to create more complex ones for the users.
5. It includes more complicated twists and turns. According to the founder, different people will use the app for different levels and so it has to be carefully prepared.
6. It plans to differentiate the content based on the objectives–like flexibility, endurance, strengthening of muscles, and on personal preferences, depending on the mood.
7. The application also analyses the pace of a person and adjusts itself making it easier for the user to adapt to yoga completely.
- It uses the neural network, and classic computer vision algorithms, with a 95% accuracy of recognition of asanas, which includes a data set of two hundred thousand images.
- These representative datasets are hard to create, particularly when it comes to finding several images of people in a particular pose.
- These images should contain different angles, and because different people in their ordinary lives have performed different yoga poses, it becomes difficult to sync it with the existing systems.
- Along with the representative dataset, another problem faced the app is the visibility of the body.
- While using Zenia, it is not recommended to keep the camera against bright light or close to a window.
Benefits of Zenia
- With Zenia capitalization, the worldwide popularity of Yoga is increased along with the help of AI augmentation.
- Not just yoga, but other domains can also make use of the concept that Zenia is applying to the practice.
- It can be used to analyze postures of different athletes and can also be used as a physiotherapy tool.
- It is not an instructor but an actual assistant which includes a real interaction between a student and the yoga instructor
- It offers more significant benefits like positioning and accuracy in the posture.
- It doesn’t matter how much Zenia denies replacing human trainers, it looks like one day AI will completely take over the fitness industry as we know it.
You can View a Trial Video of the Application
Enhancements in Zenia
- The application will only become more robust in the future. Zenia plans to track each progress by graphical illustrations.
- The app will also have social media sharing to share their progress with friends and family.
- The yoga instructors will be able to add their program and check the correctness of the asanas in a semi-automatic mode to give recommendations for improvement.
- It says that the application is not an instructor but an actual assistant which includes a real interaction between a student and the yoga instructor.