Wheel Chair Accessibility Detection

Inspiration
During my (Umair Iqbal) pilgrimage to Makkah, I pushed my grandmother's wheelchair throughout the entire journey. It was incredibly challenging, and it made me realize that accessibility needs to be easier and more intuitive. This experience sparked the idea for a solution that could revolutionize how we approach accessibility.

What it Does
Our solution helps analyze areas beforehand with just a picture, allowing you to assess whether a location is accessible or not. This enables individuals to make informed decisions, empowering them to be more self-reliant in navigating spaces and reducing uncertainties.

How We Built It
We used the Gemini API to kickstart our initial AI model. Then, we trained the model with a variety of images using Vertex AI to fine-tune its accuracy in Google AI Studio. This process enabled us to create a robust and reliable tool for accessibility assessment.

Challenges We Ran Into
During development, we faced challenges with the AI model's precision, model overloading, and issues with API connectivity. However, we overcame these obstacles through persistence and constant iterations.

Accomplishments We're Proud Of
We are proud to have built a solution that makes accessibility easier for everyone. By automating the process of analyzing accessibility features, we have streamlined the decision-making process for users and created a reliable tool for better inclusivity.

What We Learned
Throughout the journey, we gained valuable experience in using the Gemini API, Google AI Studio, and how to fine-tune AI models. The learning curve was steep, but it has been incredibly rewarding to see our idea come to life.

What's Next for Wheelchair Accessibility Detection
Our next step is to integrate the accessibility detection system with mapping platforms. By connecting it to 3D maps, we envision a future where users can view the accessibility of their destination in real-time, enabling them to make better, more informed decisions as they navigate the world.