Inspiration
Our team is comprised up of car enthusiasts who care deeply about performance, safety, and value of the only part of the car that connects the driver to the ground. We’ve always noticed that many drivers either go for the cheapest tires available or spend too much on overpriced ones, without finding the right balance between budget and safety. That frustration inspired us to create a smarter, more transparent way to help people make informed tire choices.
What it does
TireTry is an AI powered platform that helps users understand their tire health and make better purchasing decisions. Users can upload a photo of their tires, and our computer vision models analyze tread depth and wear to estimate remaining life, allowing us to classify the tires. The conversational AI platform recommends the best replacement options if needed based on driving style, budget, and vehicle type, or provides tips in general to get the best use of a set of tires. Users can even test how different tires perform in simulations under various road and weather conditions.
How we built it
We developed a computer vision model that processes tire images to identify tread patterns and wear indicators. The backend integrates AI driven recommendations powered by curated tire data and performance metrics. The frontend was built with modern web frameworks to provide a clean, easy to use experience that feels approachable for any driver.
Challenges we ran into
One major challenge was ensuring the computer vision models could accurately assess tire wear from photos taken in inconsistent lighting or angles. We also faced difficulties gathering reliable tire performance data for the recommendation engine and fine tuning the simulation to reflect realistic driving conditions.
Accomplishments that we're proud of
We’re proud of creating a working prototype that not only detects tire condition but also provides practical, personalized recommendations. Seeing our AI successfully guide users toward safer and more cost effective choices felt incredibly rewarding. Another accomplishment we are proud of is when we were faced to work with government road data. It was difficult to compare different roads because the data used mile markers for start and end points instead of latitude and longitude coordinates. We solved this by integrating a Maps API that filters and maps road data within a 20-mile radius of the user’s location, allowing for more accurate comparisons and analysis.
What we learned
We learned the essentiality of high quality data and precise image processing are for building AI tools. Carefully training data required patience and skill. We also gained valuable insight into how users interact with technology, and how technology interacts with users.
What's next for TireTry
We plan to enhance our computer vision accuracy, expand our tire database, and collaborate with tire retailers to offer instant purchasing options directly from the platform. We envision a future in which we can launch a mobile app that lets users keep track of their tire health, gaining insights from visual and location data.
Log in or sign up for Devpost to join the conversation.