sevarG

Spaces for all abilities, made easy.

Comment

Inspiration

We wanted to provide structural improvement suggestions for architecture to strive towards accessibility for people with disabilities. SevarG was inspired by Michael Graves, a disabled architect and designer renowned for his skills in architecture. SevarG is more than just a website that offers consulting. SevarG is a modern technology that promotes benefiting the needs of the disabled community through architectural improvement consulting.

What it does

Through SevarG's website, the user provides an image that must be an architecture. SevarG utilizes the Gemini Api to observe, input, and determine if said architecture is accessible to people with disabilities or not. Then, based on said outputs, the AI prompts a display to list out strengths and liabilities, highlighting them with visual boxes, and is designed to properly suggest a shopping list specifically from Lowe's Home Improvements, providing links and insight into different products the user could buy to help implement the recommended designs.

How we built it

We created our frontend through CSS & HTML. We utilized 30 Seconds of Code to incorporate interesting designs into our website. Our backend was created using the Google Gemini API. Claude and FastAPI were used to help with structuring code as well as helping us lay out our design process. We want to use Roboflow and incorporate it through computer vision in the future. We have our own data that we specifically trained and wish to include.

Challenges we ran into

We ran into a multitude of challenges stemming from the backend. The frontend was self-explanatory and ensured we stylized the website to our satisfaction. The backend had multiple problems, specifically involving the connection between the frontend and backend. The backend handles operations regarding the API; it has a variety of functions defining how Gemini should work when prompted with an image. It handles all background features (preparing a Lowe's list, recommendations, and the computer vision image). We ran out of time involving Roboflow, as the implementation would've taken longer than our time constraint. We decided to use this as our plan and incorporate Roboflow's computer vision into our project in the near future.

Accomplishments that we're proud of

We're proud to be able to host the website properly, have our backend working and connected to our frontend, and have our idea come to life.

What we learned

We learned through experience in this project to understand many of the issues that disabled individuals face in their daily lives. This has really affected us and further motivated us to continue developing severe in the future. Technically speaking, we learned API implementations, unique features of web development, we learned how computer vision algorithms work, and we learned further how machine learning algorithms operate through object detection experience utilizing yolo and Roboflow.

What's next for sevarG

We have multiple improvements we'd like to store for this project, as well as timeframes for when we'd like to push them out.

Immediate Improvements:

  • Implement user accounts and saved projects
  • Improve model accuracy with more training data and fully incorporate Roboflow's data we created into our computer vision
  • Add more marketplaces to pull from to give contractors flexibility with their purchases.

Secondary Improvements:

  • Mobile app development - push out a mobile app for SevarG to make it easier to access and utilize within the community.
  • Machine learning from user feedback
  • More functionality within the website

Built With

Share this project:

Updates