Table 4 Analysis the classification results with existing models.

From: Retinal imaging based glaucoma detection using modified pelican optimization based extreme learning machine

Existing methods

Acc(%)

Explainability

ACRIMA

RIM-One

G1020

ORIGA

TIA-Net (SOD+Attention)72

NA

–

–

–

85.70

2D-FBSE-EWT73

NA

–

–

–

91.01

SMOTE + RF74

NA

–

–

–

78.30

SMOTE + RF74

NA

–

–

–

82.80

HOG +SVM39

NA

–

–

83.32

–

HOG + PNN39

NA

–

–

87.92

–

HOG + RNN39

NA

–

–

85.72

–

SS-SQ-VDM+SVM18

NA

92.67

–

–

–

LS-SVM19

NA

84.95

–

–

–

QB-VMD20

NA

–

86.13

–

–

EWTDWT+LS-SVM21

NA

83.57

–

–

–

CVMD +SVM22

NA

–

89.18

–

–

CNN75

CAM

–

–

–

93.5

CNN76

NA

–

90.51

–

–

ResNet-5077

NA

–

–

–

92.59

Proposed model

VG, GBP, IG, GIG, SG, GCAM , GGCAM

91.45

92.43

93.25

96.75

  1. Acc Accuracy; VG Vanilla Gradients; GBP Guided Backpropagation; IG Integrated Gradients; GIG Guided Integrated Gradients; SmoothGrad; Gradient-weighted Class Activation Mapping (GCAM); Guided Grad-CAM (GGCAM)
Back to article page