Skip to main content

Table 2 Accuracy and precision of deep learning models trained from scratch for spatial resolution and geometric distortion

From: Deep Learning Approaches for Automatic Quality Assurance of Magnetic Resonance Images Using ACR Phantom

 

Spatial resolution

Geometric distortion

 

Accuracy

Precision

Accuracy

Precision

VGG16

0.46

0.6

0.49

0.6

VGG19

0.55

0.45

0.52

0.63

Resenet50

0.98

0.55

0.9

0.61

InceptionV3

0.99

0.6

0.97

0.7

EfficinetNetB0

0.97

0.57

0.95

0.61

EfficinetNetB5

0.91

0.59

0.93

0.7