Like every organization on earth we use cookies. We use cookies to analyze our product usage. We don't use cookies for commercial purposes. Learn More

AI Stories - Working out safely during covid with a fitness app super-powered by AI

In this third episode of AI Stories, Geoffrey from LabelFlow interviews Pavan Kunchala, R&D Computer Vision Engineer at FS Studio. In his free time, Pavan builds a fitness app super-powered by AI to identify and track poses.

February 2020: that’s the time when Covid locked the world inside, stuck in their homes. Pavan Kunchala, based in Hyderabad-India, is one of those people that did not have a personal fitness center in his place to continue working out properly (like probably 99% of the world population;)).

3 years ago, while studying for his computer science degree, Pavan discovered machine learning and computer vision. He then started building a fitness app for fun. The goal was initially to automatically count the number of exercises using a PC and a camera. His smart fitness app was born.


This app is for now a solo project and Pavan is building it in his free time. In order to improve the product, he benefited from a solid community of close friends and fans on Linkedin, this helped him to gather a large enough amount of feedback to improve it. Pavan is now looking for people willing to help him to accelerate a potential go to market.


Many things are still to be built: app robustification, new tracking poses, features to give recommendations on better poses, build a website, reinforce datasets for its models, etc. The end goal would be to build a whole workout platform.

As part of AI Stories Episode 3, a web series dedicated to education around Artificial Intelligence aiming to share the good vibes on AI and how it impacts our lives and our planet, Geoffrey from LabelFlow got the opportunity to exchange with Pavan:

According to KVB research The global fitness app market size was valued at USD 4.4 billion in 2020 and is expected to expand at a compound annual growth rate (CAGR) of 21.6% from 2021 to 2028. Due to covid, the global downloads of fitness and health apps increased by 46% during this period.

Pavan’s core app is built using Streamlit to turn data scripts into shareable web apps in Python. To estimate position and track poses, the app is powered by MediaPipe Pose library from Google. MediaPipe Pose is a machine learning solution for high-fidelity body pose tracking, inferring 33 3D landmarks and background segmentation mask on the whole body from RGB video frames using Blazepose


The landmark model in MediaPipe Pose predicts the location of 33 pose landmarks

We will make sure to let you know if Pavan’s promising project is made public, It is likely that there will be new cool features when it releases! If you want to connect with Pavan, you can ping him here

At the moment you can also test other leader apps on the market such a Strava, Jefit, HealthifyMe providing live and interactive workouts or Fitness + from Apple

Featured Articles

View all articles



Get news about our product and releases

© 2022 LabelFlow, All rights reserved.