A multi-task network predicts degradation patterns and a multiplicative Global Sensor Health Index from RGB images to provide early warnings of camera failure in autonomous driving before downstream detection degrades.
Making the case for safety of machine learning in highly automated driving
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Safety-Critical Camera Reliability Monitoring for ADAS via Degradation-Aware Uncertainty Pattern Analysis
A multi-task network predicts degradation patterns and a multiplicative Global Sensor Health Index from RGB images to provide early warnings of camera failure in autonomous driving before downstream detection degrades.