Zenia - AI Yoga Teacher

Your AI virtual yoga assistant 🧘

#3 Product of the DayDecember 03, 2019
Meet Zenia, the first AI-powered yoga assistant. It allows practicing yoga safely whenever you feel like it. Through the front camera, the app recognizes movements of the major joints and gives real-time comments on the quality of practice.
Discussion
Would you recommend this product?
31 Reviews5.0/5
Great job! Waiting for Android version!
Upvote (20)Share
As a yogi, I'm excited by Zenia for the real-time feedback it gives on your yoga posturing. This will be essential for an improved practice and preventing injuries.
Upvote (21)Share
Hey Hunters, I’m Alexey, the founder of Zenia. Zenia is the first virtual yoga assistant. As an engineer and an avid yoga practitioner, I wanted to combine my interests and create a product that allows practicing safely at home. The main advantage of Zenia is a personalized approach to each student. We aimed to recreate the intimate experience of yoga practice with the instructor; the app adapts to the pace, monitors progress and gives feedback on the performance. Zenia’s core technology is based on algorithms of computer vision and neural networks. Through the front camera, the app recognizes movements of 16 major joints and within seconds gives feedback. To create an accurate algorithm, we have analyzed tens of thousands of images of various asanas under the guidance of professional yoga instructors. The confidentiality of users’ data is one of our core values. Although the frontal camera is used to recognize movements, it doesn’t store images or record video format, as well as does not save anything to a server. The current version of the app features 4 guided courses and 14 sequences of different intensity and length. We’d love to hear your feedback! P.S. We offer the first month free with full access to all courses and sequences for registered users.
Upvote (20)Share
@aleksei_kurov hey Alexsei, do you say Zenia does not use any cloud SDK for recognition?
Hi, I've been working on computer vision for Zenia. If you have any questions about the tech, I'll be happy to answer.
Upvote (14)Share
@cubbeewan Really cool. Is tech home made or using any underlying SDK?
Upvote (3)Share
@cubbeewan Hi, I am a data scientist. I am fascinated by the tech you guys have built. It is really solid. How much time did it take to build the app ?
@happygiantar Hi, model inference is done with CoreML, all pre and post processing is custom.
@remidi Hi, it's hard to tell exact numbers, because we've worked on the tech on and off for a couple of years, but I would say about 8 months.
@cubbeewan That's cool. Even I am trying build something using pose estimation and I'm currently in the learning phase. It would be great to have chat with you. I have sent a invite in linkedin.
Great work, guys!
Upvote (10)Share
@ilya_lesyn Thank you!
Upvote (1)Share