CBAM-augmented EfficientNet-B3 with evidential deep learning classifies wildfire smoke severity into three levels on 16k satellite patches, reports 93.8% weighted accuracy, and decomposes epistemic and aleatoric uncertainty without Monte Carlo sampling.
Title resolution pending
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
-
Uncertainty-Aware Wildfire Smoke Density Classification from Satellite Imagery via CBAM-Augmented EfficientNet with Evidential Deep Learning
CBAM-augmented EfficientNet-B3 with evidential deep learning classifies wildfire smoke severity into three levels on 16k satellite patches, reports 93.8% weighted accuracy, and decomposes epistemic and aleatoric uncertainty without Monte Carlo sampling.