Table 1 Final network layer correlations with shape and category for Set A and B. Maximum values across the four networks are highlighted in bold.

From: Orthogonal Representations of Object Shape and Category in Deep Convolutional Neural Networks and Human Visual Cortex

Network

Set A

Set B

Shape

Category

Shape

Category

CaffeNet

0.5965

0.3737

0.3639

0.2394

VGG-19

0.2888

0.3556

0.2786

0.3605

GoogLeNet

0.4375

0.4247

0.3125

0.3569

ResNet50

0.4978

0.2209

0.3159

0.3652

Back to article page