Abstract
Current waste and recycling programs are suffering from consumer confusion, which causes contamination. To prevent contamination, we developed an Application Programming Interface (API) that uses object recognition and the consumer's location information to identify recyclable materials and inform the consumer. Our API can be used to educate consumers on how to properly handle their materials according to local laws and materials accepted in specific collection programs. The API delivers proper materials handling responses combining computer vision and database analytics. As an example use‐case for our API, we collected a custom dataset of different mixed plastics and electronics and fine‐tuned several existing object recognition models using our dataset. We assess the performance of each model on our custom dataset with different metrics. We describe all the components of our API, which provides a blueprint for deploying existing state‐of‐the‐art computer vision models in a commercial application.