Bin contamination is a huge problem
Errors in waste disposal are not only missed opportunities to recycle or compost, but also lead to the contamination of entire recycling bins. Often, an entire waste bin can end up into a landfill due to a single error leading to contamination.
Data from the National Waste and Recycling Association shows that human confusion in the disposal of waste in waste bins results in nearly 25% of recyclables getting contaminated, diverting materials that could be recycled in our landfill
When compostable materials such as food scraps gets into a landfill, it is compacted down and covered. This removes the oxygen and causes it to break down into an anaerobic process. Eventually, this releases methane, a greenhouse gas that is 25 times more potent than CO2 in warming the earth
Existing approaches are inaccessible, expensive, slow, and inaccurate
Confusing signs found at waste bins are hard to understand and are incomplete.
Expensive and hard-to-scale “smart bins” at recycling center are unable to prevent the bin contamination that occurred at the time of disposal due to humans.
Current machine learning research suffers from low accuracy (ranging from 22% to low 70%) or is too slow for real-time usage within a mobile application.
None of these approaches target compost classification. Accurate compostable disposal is important because when compost ends in landfills → methane emissions
Effective solution needs to be fast, accurate, and accessible at the time of waste disposal
DeepWaste utilizes the recent breakthroughs in AI for image recognition and the increased computational power on everyday cell phones to provide a novel approach for waste classification that is accurate, low-cost, fast, and accessible.