Skip to main content
. 2022 Apr 25;13:829027. doi: 10.3389/fmicb.2022.829027

Table 2.

Parameter setting of EMDS-6 feature classification using SVM.

Feature Kernel C DFS Tol Max iter
LBP rbf 50,000 ovr 1e-3 -1
GLCM rbf 10,000 ovr 1e-3 -1
HOG rbf 1,000 ovr 1e-3 -1
HSV rbf 100 ovr 1e-3 -1
Geo rbf 2,000,000 ovr 1e-3 -1
Hu rbf 100,000 ovr 1e-3 -1
RGB rbf 20 ovr 1e-3 -1

C, penalty coefficient; DFS, decision function shape; tol, the error value of stopping training; Geo, geometric features.