Table 8 The suggested ViT-GAN-SVM framework cross-modal generalizability.

From: Enhanced classification prostate cancer based on generative adversarial networks and integrated deep learning with vision transformer models

Dataset

Modality

Accuracy (%)

Sensitivity (%)

F1-Score (%)

PSNR (dB)

SSIM

Prostate DWI (original)

DWI

99.78 ± 0.12

99.85 ± 0.08

99.85 ± 0.09

45.87 ± 0.33

0.980 ± 0.002

ProstateX

T2w + ADC

98.64 ± 0.27

98.63 ± 0.29

98.65 ± 0.28

44.12 ± 0.32

0.972 ± 0.004

BraTS 2021

Multi-MRI

97.47 ± 0.35

98.23 ± 0.40

97.64 ± 0.37

43.68 ± 0.50

0.968 ± 0.005

LIDC-IDRI

CT

96.21 ± 0.43

96.34 ± 0.44

96.53 ± 0.41

42.91 ± 0.58

0.961 ± 0.006

Back to article page