Journal of Remote Sensing / 2021 / Article / Tab 8

Research Article

Shearlet-Based Structure-Aware Filtering for Hyperspectral and LiDAR Data Classification

Table 8

Classification performance using Raw-H, Raw, 3D-CNN, miniGCN, SAE-LR, NMFL, EPCA, GGF, EMAP, OTVCA, and ShearSAF for the MUUFL Gulfport dataset with five labeled samples per class as the training set.

ClassRaw-HRaw3D-CNNMiniGCNSAE-LRNMFLEPCAGGFEMAPOTVCAShearSAF

C167.0670.4583.1579.9868.0969.7763.2761.1070.1375.0583.52
C270.8571.4468.8490.5858.1167.4270.5472.0673.2653.3272.54
C342.4742.9537.4338.6030.1650.1836.4752.5240.6832.4958.26
C447.3347.3861.8159.9661.7368.5446.7149.7261.8463.7865.74
C569.6372.6866.4281.1357.1273.7576.0479.6780.5854.0282.70
C684.2486.1561.0394.4165.2198.4097.9996.7896.7599.8787.55
C762.3868.8671.6276.1955.7372.2982.6076.6477.6974.9989.25
C832.2339.1774.5478.4179.2859.2773.5952.8170.9066.8064.08
C942.4643.8338.0640.0829.7045.4739.5745.6239.8734.4048.61
C1061.0761.6449.2176.4040.8284.7853.3361.4559.7369.4061.72
C1190.3790.6561.2372.3042.8674.3193.0988.9094.3396.4989.96
OA59.2362.3270.2868.6960.4066.1163.2562.5368.0564.6072.30
Kappa0.500.530.620.620.510.580.550.540.600.560.65