General description : A sports companion application designed to support users in achieving their fitness goals safely and effectively. Our goal is to provide a tool that not only helps beginners develop good exercise habits but also serves as a valuable resource for athletes of all levels looking to monitor and improve their performance. Overall, our app aims to prevent injuries by promoting proper form and technique, set personalized workouts for continuous motivation, and track progress in an interactive way. App features :
- Using computer vision and machine learning, the app analyzes workout sessions in real-time through the user’s device camera. It provides instant feedback to ensure proper form and technique, such as correcting shoulder position during a plank but also helps counting repetitions and keeping track of key performance metrics (repetition for Push ups, squats etc and time for chair, planck etc).
- The app offers a running feature that allows to track their path and save it. By using GPS data, the app also measures pace and distance in real-time to help users achieve their target times and improve their endurance.
- Users can view their workout history in a calendar format. This feature allows users to track their progress over time, visualize their consistency, and stay motivated to reach their fitness goals.
- The app visualizes progress through graphs and charts, highlighting improvement in areas such as strength or endurance.
- The app offers workout creation for both bodyweight and yoga. And then the user can follow the flow of his workout. While doing the workout he can film himself and compare his position with a coach video.
- When the user finishes a workout he can save it, so that he can import it back if he wants.
Technical approach : We use Google MediaPipe to capture 3D body joints positions and compute angles for accurate form correction. For moving exercises (squats, push-ups…) , the app use this data to track movement paths and provide feedback if the user happens to deviate from the correct form. For static exercises (planks…), the app makes sure that all critical angles are within safe ranges. The app can also count repetition for example for push ups.
External Resources :
- Our app uses Google Firebase to store user data such as workout history, progress, friends, or even preferences.
- User Support: Users can create accounts using Android’s built-in Google authentication to track their progress.
- Sensor use: We use the camera for workout analysis as well as GPS for running plans.
- Offline mode: Without internet connectivity, users are still able to perform cached workouts. They can also still use the ML feature.
We have a Figma project for the app design and navigation as well as the architecture diagram. Here is the link to the project.
Here is the link to the architecture diagram of the app. link