Fig. 9
From: Multiscale superpixel depth feature extraction for hyperspectral image classification
Salinas dataset. (a) Ground truth. Classification maps obtained by different classification methods. (b) 2DCNN, OA = 84.22%, (c) MCMs, OA = 90.32%, (d) HybridSN, OA = 90.78%, (e) SSFTT, OA = 93.99%, (f) CEGCN, OA = 92.79%, (g) MSSGU, OA = 94.91%, (h) SBD, OA = 94.30%, (i) SHDA, OA = 92.07%, (j) AMGCFN, OA = 92.91%, (k) MSDFE, OA = 97.85%, (l) Labels.
