Citizen Science Mobile Apps with Machine Learning for Recyclable Objects
Chelsea S. Yeh, Fahim Hasan Khan · 2022
We describe the application of citizen science mobile applications (also known as “apps ”) with integrated machine learning for the real-time identification of recyclable objects. Citizen science mobile apps enable the collection and dissemination of scientific data to a broad non-scientific community by utilizing existing sensors and devices on ubiquitous mobile phones. In this research, we created mobile apps with integrated pre-trained machine-learning models to identify common recyclable objects captured with mobile phone cameras. The models were trained with images of recyclables on a server and loaded into the mobile phones for client-side identification. The mobile app has applications for quickly identifying recyclable items that have been discarded, providing recycling regulations and information based on location, and gathering data regarding the density and frequency of improperly discarded recyclables.