Pro Weightlifting

A mobile app that uses computer vision to make individual weightlifting workouts safe and effective

Business Analysis
DevOps services
Mobile Development
MVP
QA
UI/UX design
Web Development

Project description

We created this advanced sports application for professional and non-professional weightlifters worldwide. The app’s core functionality is the computer vision-driven feature that recognizes the barbell movement trajectory and compares it with the correct one to ensure that every workout is safe for the user and helps them reach good results. The app considers the user’s results after each workout and calculates a new workload for the next training session. Additionally, users can make notes on each training and write anything related to it – physical and emotional condition, observations, thoughts, etc. Another core functionality is nutrition recommendations, and it helps boost the efficiency of training with an individual diet.

Industry:	Sports
Platforms:	Mobile
Role:	Software Developer, IT Consultant
Tech Stack:	Python, Node.js, React Native

Problem introduction

Our client is a world-class professional weightlifter. He wanted to create a mobile application that would allow weightlifting professionals and enthusiasts to have safe and effective workouts at home. Because of the pandemic and inability to workout in a gym, under the close attention of a professional trainer, people interested in weightlifting faced challenges to get good-quality individual training. A mobile application that includes artificial intelligence can make this problem go away. Our team was primarily responsible for technological implementation. Meanwhile, sports, nutrition, and rehabilitation experts created recommendations for eating and training so that our system can use them to develop effective individual plans.

Main goals:

Cover iOS and Android with one cross-platform mobile app
Integrate computer vision solutions to the platform
Build a system that automatically generates safe and effective individual nutrition and training programs
Ensure high-quality calculation results for nutrition based on height, weight, age, and gender
Create an informative landing page

Tech Challenges

The implementation of computer vision functionality was challenging, especially making the trajectory line move naturally on the video. We strive to make all our web and mobile solutions both functional and aesthetically pleasing. So, it took some time for our Python developer to create a frame-by-frame mask movement that would be smooth and go in sync with the body and barbell movement speed. We used different methods and, with the help of a unique filter, achieved the perfect result you can see on the app.

Core Functionality:

Computer vision-based Barbell Path feature
Training programs with exercises
Nutrition plans for different goals (gaining, losing, and maintaining weight)
Video downloading and recording

Solution

The result of our work is a unique mobile app for iOS and Android that includes functionality for effective training and nutrition. The app covers all essential training needs of weightlifting professionals and enthusiasts – workout technique, individual training and nutrition recommendations, notes on each workout, etc.

The core functionality is the Barbell Path. A user makes a short video of working with a barbell; then, the system uses computer vision to recognize the body and the barbell. The app builds the movement trajectory line on the video, compares it with the perfect one, and shows whether the user works with the barbell correctly or has some technical errors that may lead to poor training results or even injuries.

The nutrition functionality is based on the recommendations of a professional nutritionist, who created diet plans for various needs like losing, gaining, and maintaining weight. The system gives calorie and product recommendations based on the user’s weight, height, age, and gender.

Business analysis showed that weightlifting is mainly popular in Brazil, the United States, Europe, India, and a few other countries, so we made the app available in 10 different languages.