Cascina Elisa Politecnico di Torino [email protected]

Pellegrino Andrea Politecnico di Torino [email protected]

Tozzi Lorenzo Politecnico di Torino [email protected]

What is Semantic Segmentation for waste sorting?

Semantic segmentation for waste sorting is a computer vision technique that helps machines understand and separate different types of waste in images or videos. It's like teaching a computer to recognize and categorize objects in a picture, but specifically for sorting trash.

Here's a simple explanation:

  1. Image Analysis: Imagine you have a photo of a messy pile of waste items, such as bottles, cans, and paper. Semantic segmentation is like giving the computer the ability to look at each pixel in the image and decide what type of waste it belongs to.
  2. Coloring the Objects: The computer assigns different colors to different types of waste. For example, it might color all the bottles green, all the cans red, and all the paper blue.
  3. Separating and Sorting: Once the computer has "colored" each pixel in the image, it can separate the waste items by their colors. This allows it to recognize and sort the trash automatically into different categories, like recycling bins for bottles and cans or a bin for paper waste.

Example of what a semantic segmentation model can do

Example of what a semantic segmentation model can do

What were our goals?