Skip to main content
. 2023 Mar 16;9(3):69. doi: 10.3390/jimaging9030069

Figure 7.

Figure 7

Comparison of UMap projection of activations of images generated by a DCGAN, and others generated by a StyleGAN, with an InceptionNet trained on ImageNet (top row), and a U-Net trained on the original dataset (bottom row). Red points: real images; blue points: generated images.