Open access
Oct 2026
Spatial–Spectral Prototype-Guided Deep SMOTE for Imbalanced Hyperspectral Image Classification
Spatial–Spectral Prototype-guided Deep SMOTE (SSPD) is proposed for imbalanced HSI classification that achieves more stable and competitive performance in terms of Overall Accuracy (OA), Average Accuracy (AA), and Kappa Coefficient (KC).
Yan Lv, Guang-Hao Kang, Yan Cao et al.
· Remote Sensing · 0 citations