Skip to main content
. 2021 Jul 5;111:107669. doi: 10.1016/j.asoc.2021.107669

Table 1.

Details of SLB methods.

SLB
model
Pre-trained
model
Input
size
Feature
layer
Feature
vector
SLB1 AlexNet 227×227×3 fc6 4096
SLB2 ResNet18 224×224×3 fc1000 1000
SLB3 ResNet18 224×224×3 pool5 512
SLB4 ResNet50 224×224×3 fc1000 1000