Inspiration
We wanted to create an easy-to-use fitness app that assists people of all levels with their fitness journeys.
What it does
EZX is a website with three features:
- AI-customized diet and workout plans.
- A chatbot named Vivo that answers any fitness-related questions.
- A video form analyzer that provides the user feedback on their exercise form
When the user first opens the website, a login screen appears. They then type in their information and submit it. After they submit it, they arrive at the main page of the website where they can see their customized diet and workout plan, ask questions make adjustments to their plans using Vivo, or upload a video of them to our form analyzer.
How we built it
-We built Vivo using the UMGPT API and made changes in Visual Studio Code.
-We built the website using react and connected the back end using PowerShell.
-We created the form analysis using scikit-learn and mediapipe's pose class. With the help of mediapipe's pose library we were able to identify and extract the coordinates of the joints easily and then we used RandomForestClassifier in Scikit-Learn to train and test the model.
Challenges we ran into
- We had to do a lot of experimentation to get Vivo to output a diet and workout plan with a consistent format.
- It was difficult to connect the front end and back end. The main difficulty was the size of the videos
- When building the form analyzer, the model was overfitting, and it wasn't efficient enough. We didn't have enough time to get a large enough database.
Accomplishments that we're proud of
We are proud that we created all of the features in only 24 hours. We created an entire website equipped with a chatbot and a form analyzer. We are especially proud of the form analyzer because it took a lot of time to train the model how to recognize incorrect form.
What we learned
We learned how to customize an API to use it for a specific purpose. We also learned how to train the model to analyze whether the exercise is in the up or down position and if the form is correct.
What's next for EZX
First, we will add additional exercises to the form analyzer feature. Right now, it only analyzes push-ups and squats but in the future, we will continue to expand the database to analyze as many different exercises as possible.
Marketing Strategy
Digital Advertising: Utilize social media, search engine marketing, and targeted online ads to reach potential users.
Content Creation: Publish informative blogs, videos, and tutorials on fitness, attracting and engaging the target audience.
Partnerships and Collaborations: Collaborate with fitness influencers, trainers, and health-related brands to increase brand visibility.
Revenue
Subscription Model: Offer tiered subscription plans with varying features and benefits.
Premium Services: specialized programs, or advanced analytics available for an extra fee.
Affiliate Partnerships: Collaborate with fitness-related brands and earn commissions through referrals.
Advertisements: Advertise fitness-related content on the website
Log in or sign up for Devpost to join the conversation.