Sensing / 2026
TIPS
Combine active thermal imaging and physics-informed learning to identify plastics.
About the project
TIPS classifies plastic types from their thermal response. A two-stage training strategy pretrains with simulated heat-transfer data and then adapts to real samples, connecting physical models with practical sensing.
What it enables
The project reports up to 100% classification accuracy for black plastics and 94.7% for white plastics in its evaluated samples.