Skip to main content

Table 7 Using different levels of convolutional bases as the level name and feature size of the feature extractor

From: Fusing pre-trained convolutional neural networks features for multi-differentiated subtypes of liver cancer on histopathological images

Numbering

Combined content

Feature length

A1

ResNet50(full convolutional basis)

2048

A2

DenseNet201(full convolutional basis)

1920

A3

VGG16(full convolutional basis)

512

A4

InceptionRenseNetV2(full convolutional basis)

1536

B1

ResNet50(conv4_block6_out)

1024

B2

DenseNet201 (conv4_block6_out)

1792

B3

VGG16 (block4_pool)

512

C1

ResNet50(conv3_block4_out)

512

C2

DenseNet201 (conv3_block12_concat)

512

C3

VGG16 (block3_pool)

256

Back to article page

BMC Medical Informatics and Decision Making

ISSN: 1472-6947

Contact us