The problem FitVerse solves

People with intellectual disabilities (ID) have high prevalence of physical inactivity and high rates of preventable health conditions, including obesity and cardiovascular disease. Increased participation in fitness and wellness activities can improve their overall health and well-being. However, many fitness and wellness organizations do not know how to effectively serve this population nor how to make an effective workout environment. When it comes to go out for workout , consistency breaks. If we do exercise alone at home, lack of proper guidance can lead to injuries. Doing any work without any reward tends to make that work quite boring. Knowing information about current health state (like: Blood pressure, SPO2 level, Pulse rate) is necessary during proper exercise.

FitVerse is specifically designed to solve all these problems in most optimum way .
We developed a single platorm using augmented and virtual reality, machine learning , open cv , and a mobile app for storing and tracking of information of daily workout routine.
The system will also store SPO2 , BP and pulse rate on raily basis during workout session via connecting to smart watch.
Proper guidance and tracking of exercises will be provided using machine learning and ar models.
System will use voice commands to guide availability of proper login facilities for users and personalized stats and proper reward system for maintaining consistency.
During workout user will be introduced to virtual world via vr- headset . It utilizes the cutting edge , immersive technology to build new workout environment at your own home.

Challenges we ran into

We were completely new to many tech stacks and modules that we used in the project. Using Voice command and building proper warning system was a major task. Collecting, recording and selecting dataset for Different posture was a tedious task.
We solved it by calling 4 different basic machine learning model and testing there accuracy score and taking the highest accuracy model for testing.
Training the model for common human errors during excercise took a lot of efforts.We trained model with proper specific poses, and removing any disparity while recording data.
Building proper environment using Unity and Blender was again challenging . We were required to solve many system related errors while scripting in Unity 3D and importing models from blender to Unity along with all the textures and animations especially while building virtual 3D trainers performing all the workout and exercises step by step for the smooth learning of user.
With all the efforts shown by our team , we came across all the odds and finally accomplished our target on time.