Remote-Sensing Image Tile-Level Annotation Based on Discriminative Features and Expressive Visual Word Descriptors

Table 5

The OA and Kappa comparison between using four different methods on the second dataset.

Methods	OA	Kappa
	0.892 6	0.810 3
OB-LDA	0.914 6	0.818 5
LEGION + SVM	0.942 4	0.835 6
Our proposed method	0.946 0	0.837 1