• It can be hard to know what trash to sort, and recycling or compost plants may not be able to tell if someone put something they shouldn't have in the recycling / compost

What it does

  • This application takes in a video feed and outputs what you should do with the item in the image (recycle, compost, garbage, etc.)

How we built it

  • We used AWS's Rekognition to process images from the video feed
  • Using the identified images, we identified some common recyclable / compost items and decided what to do with them (turns out you shouldn't tell users to put people in the garbage)

Challenges we ran into

  • Image recognition isn't always the most accurate - it largely depends on camera angle and how stationary the video is

Accomplishments that we're proud of

  • It works on a stationary video feed for some common recyclable / compost items

What we learned

  • How to use AWS Rekognition

What's next for Sortify

  • making the algorithm more sophisticated by running terms through a smart algorithm that will match the word to its synonyms (ex. the algorithm shouldn't need to sort for both paper and cardboard, the smart algorithm will see "paper" and add "cardboard" automatically)
  • matching more recyclable / compost items
  • not matching with things that shouldn't be thrown out
Share this project: