Skip to main content

Table 5 Comparison results of multi-modal models with (without) ACME

From: An adaptive multi-modal hybrid model for classifying thyroid nodules by combining ultrasound and infrared thermal images

Model

ACC

PRE

SEN

SPE

F1

F2

ResNet w/o AMCE

0.8444

0.9000

0.7472

0.9283

0.8165

0.8854

ResNet w/ AMCE

0.9406

0.9358

0.9358

0.9446

0.9358

0.9428

ViT w/o AMCE

0.8357

0.8178

0.8302

0.8404

0.8240

0.8383

ViT w/ AMCE

0.9476

0.9242

0.9660

0.9316

0.9446

0.9383

Hybrid w/o AMCE

0.8636

0.8880

0.8075

0.9121

0.8458

0.8891

AmmH

0.9738

0.9699

0.9736

0.9739

0.9717

0.9738

  1. Bold values indicate the best results achieved in each indicator
  2. “w/” indicates “with”; “w/o” indicates “without”; “ACME” indicates adaptive cross-modal encoder
Back to article page

BMC Bioinformatics

ISSN: 1471-2105

Contact us