Table 14 Quantitative explainability assessment results.

From: Transfer learning with fuzzy decision support for multi-class lung disease classification: performance analysis of pre-trained CNN models

Model architecture

Faithfulness score

Comprehensibility index

Clinical consistency score

Average explanation time (ms)

VGG16 + fuzzy

0.847 ± 0.023

0.781 ± 0.031

0.792 ± 0.028

145.3 ± 12.7

VGG19 + fuzzy

0.862 ± 0.019

0.798 ± 0.027

0.816 ± 0.024

158.7 ± 15.2

ResNet50 + fuzzy

0.891 ± 0.015

0.823 ± 0.022

0.847 ± 0.019

132.4 ± 10.8

Back to article page