According to the World Bank Report, "Humankind currently produces two billion tons of waste per year between 7.6 billion people" This is very damaging and can be reduced through many different implementations. We hope to combat part of this problem by removing the factor of improper waste disposal. Have you ever been unsure of how or where to dispose of your garbage? ZeroWaste will tell you exactly how it should be disposed of with a simple scan using your smartphone's camera. Information from an uploaded photo will provide the user with the proper disposal protocols based on the object in focus on their provided image. Improper garbage disposal is a huge problem for the ecosystem and ZeroWaste would like to reduce the impact of environmental damage. For ZeroWaste we have used different technologies and allowed them to work together for our application processes. We have used Core ML to define our garbage sorting model which will take a user's image in order to define what kind of waste an object is and how it should be disposed. This information is sent to the user seconds after sending their image for processing. Furthermore, we have made use of Google Firebase to store information on region-based rules as well as our garbage categories. We believe that Firebase is a reliable cloud server for storing this information and it allows the application to run at a smoother pace since the app will take up less space on the users' device. Presently, our application is working with 10 garbage categories only, we wanted to make sure that our model will run smoothly before adding more categories. We wanted our model to appropriately address region-based rules and work properly at recognizing garbage categories before we moved on to improve our features and overall UX of ZeroWaste. In the future we hope to hash out these smaller issues and build on our app so that it is able to handle more garbage categories and work in a wider range of regions. We would like for our users to be confident that they are working with a reliable application that they can trust.


This was our first idea for the capstone project. Zoe introduced the idea to us in the first semester of capstone, we noticed a lot of information around campus about being mindful of personal waste and proper disposal of garbage seemed to be one of sheridans missions for its students. We thought that this is a simple yet important idea, that can be used to push the importance of proper waste disposal. To start the project we did research on different model types to see which model would suit our idea the best. We initially wanted to use Microsoft Azure's 'Computer Vision' model. We found that this would be more complicated as it did not directly suit our needs and we found that integrating different features/frameworks using this model would be more difficult than we thought. After exploring more modelling options, we decided to use Apple's Core ML model to support our idea since we would be able to seamlessly include other features and frameworks to our app through the Apple environment. Last semester, our team created the simple UI and trained a test ML model. We used SQLite as the database to store user accounts. However, it can only recognize the plastic bottle. Besides, the ML model is locally on iOS devices. While our application was slowly coming together, it was still very new and not effective for different materials. To grow this part of our application, we trained our model with five times more images and added a cloud-based database to store information rather than locally on SQLite. These changes improved our model recognition and speed. Currently our app recognizes 10 different garbage categories and is able to change feedback based on regional rules. We are proud to see the growth of ZeroWaste and hope to expand even more as we continue to develop.


  • ZeroWaste can recognize the objects type
  • ZeroWaste can classify which garbage bin the garbage belongs in
  • ZeroWaste can classify if your region has different garbage rules
  • ZeroWaste can show your current location and surrounding waste facilities
  • ZeroWaste can store user feedback for model recognition and rule updates
  • ZeroWaste can show live AR handbook to educate about waste disposal


Zero Waste Pitch


About ZeroWaste (Group 6)

Our team at ZeroWaste has developed with skills acquired through our time as students at sheridan college while working through this project in our final year of study. We beleive that our dedication to thie project has allowed us to learn new features that we can add to ZeroWaste. We have been developing using Core ML, Google Firebase, frameworks provided by Apple Developer such as Mapkit and ARkit. We have been able to put these together through our developement to produce a candidate product that is now able to be improved and polished through regular updates. ZeroWaste would like to be the simple answer for someone who is unsure of their garbage rules or is concerned with properly disposing their garbage. We would also like people to be more considerate of their waste and learn a bit about waste disposal. We have incorporated a few features to support this goal. We hope to educate communities to become more educated on this problem and improve their quality of life.

ZeroWaste Credits

Xuanchen Liu
Developer, Database Designer
Haoyue Wang
Developer, UI/UX Designer
Saam Haghighat
Developer, Project Manager