May 2019
Intermediate to advanced
272 pages
7h 19m
English
The Discriminator architecture is a map from a multidimensional space where images live to a much smaller dimensional space: the number line. This number line will be used to compute the probability that the input image is fake or real. Following the trend in computer vision, the Discriminator architecture uses convolution layers to learn this projection. Our first Discriminator implementation is based on a fully convolutional architecture called DCGAN. We provide DCGAN's discriminator diagram below.

Read now
Unlock full access