Compostable

Inspiration

My inspiration for this project cam from my grandma, who always told me to eat everything on my plate and not waste anything. It got me thinking about the food waste I see at school and the implications of it to society as a whole.

What it does

Compostable takes in images and classifies it as biodegradable or non-biodegradable.

How we built it

I used Lobe.ai to train the model due to not having enough time to fully learn tensorflow. I also used a dataset from Kaggle

Challenges we ran into

I did not have any experience with ML before this as most of my coding experience is in robotics control programming and usaco. This was extremely challenging for me but I overall think this experience was very positive and ML is very cool. I was originally planning to learn TensorFlow for this project but due to the lack of time I had to use Lobe. I also had a hard time finding datasets as the datasets i found where either lists of flicker links or incomplete. However, I eventually found one in a zip on kaggle.

google drive link if youtube does not process: https://drive.google.com/file/d/1dg8H02Qe-4YRcUy-_ezH58H3MZ-x_p5N/view?usp=sharing

Accomplishments that we're proud of

I proud of pulling off this project as a solo team and learning ML and the training process along the way!

What we learned

I learned so so much ML from this and I'm excited for the next hackathon!

What's next for Compostable

• Added support for classifying food-like toys/objects
• develop it into an app
• add support for robotic arm to do object tracking on the trash and pick it up to achieve full autonomy

Built With

• lobe