A Robust Context-Based Deep Learning Approach for Highly Imbalanced Hyperspectral Classification

Table 6

Average accuracy score during the testing phase (model-prediction).

Models	Clean datasets		Noisy datasets
	Pavia Centre (%)	Pavia University (%)	Pavia Centre w/noise (%)	Pavia University w/noise (%)
DCN	88.92	83.37	68.98	64.79
PCA + DCN	88.52	79.76	40.62	33.93
Context-based DCN	89.88	85.02	88.01	81.99