Table 6 Feature comparison analysis.

From: Modified ShuffleNet trained on gradient pattern and shape-based features for lung cancer classification with improved M-SegNet segmentation

Methods

With only improved LGP

With only shape based feature

Without feature extraction

Proposed

Accuracy

0.929127

0.885941

0.907534

0.946581

Sensitivity

0.925304

0.885404

0.905354

0.939024

Specificity

0.933363

0.886533

0.909948

0.954955

Precision

0.938455

0.892028

0.915241

0.958506

f_measure

0.931823

0.888704

0.910263

0.948665

mcc

0.863136

0.820317

0.841726

0.893203

npv

0.91915

0.879345

0.899247

0.933921

fpr

0.066637

0.113467

0.090052

0.045045

fnr

0.074696

0.114596

0.094646

0.060976

Back to article page