Skip to main content

Table 3 Classification performance with weight balancing technique (WB), oversampling technique (OS), and mosaic-based soft labeling (MBS)

From: Enhancing deep learning classification performance of tongue lesions in imbalanced data: mosaic-based soft labeling with curriculum learning

 

precision

recall

\({{\varvec{F}}}_{1}\) score

 

WB

OS

MBS

WB

OS

MBS

WB

OS

MBS

cancer

0.8293

0.8250

0.8500

0.9189

0.8919

0.9189

0.8718

0.8571

0.8831

OPMD

0.6875

0.6364

0.8333

0.7857

0.5000

0.7143

0.7333

0.5600

0.7692

noncancerous lesion

0.9919

0.9612

0.9844

0.9457

0.9612

0.9767

0.9683

0.9612

0.9805

weighted average

0.9348

0.9080

0.9450

0.9278

0.9111

0.9444

0.9302

0.9086

0.9441

macro average

0.8362

0.8075

0.8892

0.8835

0.7844

0.8700

0.8578

0.7928

0.8776

Back to article page

BMC Oral Health

ISSN: 1472-6831

Contact us