Garbage Detection Algorithm Based on Deep Learning
Challa Naresh,
B. Gayathri,
B Pranavi,
G Laxmi Prasanna
Enabling automation technologies such as garbage detection can drastically improve waste management.
Understanding the importance of waste disposal management, this paper provides a detailed project on
the deep learning approaches for detecting garbage, specifically the Mask R-CNN model through the
Tensor Flow framework. Given that Mask R-CNN is the extension to Faster R-CNN, instead of just
object detection, it offers instance segmentation too. In this case, the image can segment garbage into
multiple types. This was modeled using COCO as the base model. The model will be able to reuse the
information from previous learnings and apply it to the new context of garbage detection. The system
has shown its potential to locate and cut rubbish even in difficult circumstances which can only be an
asset in the automation of waste disposal systems.
