EcoScan

• 

  Title Screen

• 

  Scan Screen

• 

  AI Screen

• 

  Recognition Screen

Inspiration

My inspiration for EcoScan came from observing the challenges people face when trying to recycle correctly. I wanted to create a solution that simplifies the process and makes recycling more accessible for everyone. In addition, I have always loved technology and programming, so I wanted to implement AI to help solve the issue.

What it does

EcoScan uses AI image classification (MIT look extension) and AI chatbot (ChatGPT) to help users determine the correct recycling process for various objects. By simply scanning an item with their phone camera, users receive instant guidance on the correct recycling procedure. In addition, there is also a speech recognition feature that allows users to use their voice to talk to the AI.

How we built it

I developed EcoScan by integrating different AI technologies. I first used the MIT App Inventor to create the app. Then for the image classification, I used the MIT look extension to recognize scanned objects. This extension was pre trained to recognize different materials and items. For the AI chatbot, I used the API from ChatGPT to provide clear, actionable recycling instructions based on the scanned object.

Challenges we ran into

One of the main challenges was implementing the AI features and creating a user-friendly interface. For example, when implementing the image classification, I often faced many bugs, such as the camera not loading or being stuck in one screen. In addition, when designing the interface, I had to ensure the elements were properly positioned with the colors, dimensions, etc.

Accomplishments that we're proud of

I am proud of creating a tool that not only simplifies recycling, but also educates users on proper waste disposal. I believe my app has a high level of accuracy on the object identification, and I believe that it can truly help others who are unsure of the correct recycling procedure.

What we learned

I learned maintaining high accuracy in AI image recognition requires continuous updates and a lot of data. Many of the objects that were incorrectly labeled was due to the lack of trained data. I also realized that educating users about proper recycling practices fosters greater environmental responsibility and encourages positive behavioral change, since when my younger brother used my app, he started being more mindful of recycling at home.

What's next for EcoScan

Next, I plan to expand EcoScan’s capabilities by incorporating more features, like language translation and plan to create my own AI model to handle even more types of materials. I also aim to integrate location-based recycling information to provide users with more specific guidance tailored to their local recycling rules.

Built With

• chatgptapi
• mit-app-inventor