A Comparative Analysis of Visual Encoding Models Based on Classification and Segmentation Task-Driven CNNs

<div>Contribution to the prediction accuracy of each layer in (a) VGG16 and (b) FCN32 networks. The ordinate represents the contribution of each layer, and the abscissa represents the five visual areas. The color bar from deep to shallow indicates the network layer from a low level to a high level.</div>

Computational and Mathematical Methods in Medicine

fig11

Figure 11

Figure 11: A Comparative Analysis of Visual Encoding Models Based on Classification and Segmentation Task-Driven CNNs 