Eco-Detective

Think Outside the Trash

Comment

Inspiration

As people have spent much more time in their houses due to the pandemic, trash buildup has become significantly more noticeable. Our team wants to reduce that trash, but also sustain the environment. The education system does not teach students whether an item is recyclable/compostable or not; furthermore, the benefits are severely underemphasized. To make up for this, our team wanted to create a mobile app that would assist people in their everyday lives amidst the pandemic and also exhibit the good they are doing by recycling/composting certain items.

What it does

This mobile application allows users to take pictures of items(up to 5) to identify if they are recyclable/compostable or not. If the object is recyclable/compostable, the application shows the benefits of recycling/composting that certain object.

How we built it

The base software we utilized to create our application was Android Studio, using Java. Our team also incorporated machine learning with the TensorFlow Lite API to help identify objects

Challenges we ran into

One of the biggest challenges our team ran into was getting boxes drawn around items that the camera was detecting. To overcome this problem, we thought of a workaround, where a new screen would be displayed with all of the recyclable and compostable objects. Additionally, we were able to allow users to choose the number of items being detected to prevent any confusion in spaces with multiple objects.

Accomplishments that we're proud of

One major accomplishment that we are extremely proud of is successfully incorporating the object detection feature of TensorFlow Lite into our application. In the past, one of our members struggled to get object detection working for his other projects, but we managed to pull it off for this application. Another accomplishment that we are proud of is how fast we were able to get the camera feature working. Taking a photo and storing it as a bitmap seemed a bit daunting initially, but we were able to overcome this task to focus on other functionalities of the application.

What we learned

We learned the basics of convolutional neural networks, deep learning, and region-based convolutional neural networks (R-CNN). This was the first time working with machine learning for most of us, so we got to absorb new techniques for our future projects. Also, our team had barely any background in Android Studio, so getting used to the interface was a huge learning experience for a majority of the members.

What's next for Eco-Detective

In regards to the application, we would like the create our own model for our algorithm. Due to the limited amount of time, we could not collect enough labeled data for recyclable and compostable objects. Thus, we resorted to using the base model offered by Google. In regards to the growth of the application, we would like to partner up with public schools and corporations to spread awareness of the benefits of recycling and composting and ultimately preserve the environment.

Built With

Share this project:

Updates