StyleGAN vs StyleGAN2 vs StyleGAN2-ADA vs StyleGAN3

In this article, I will compare and show you the evolution of StyleGAN, StyleGAN2, StyleGAN2-ADA, and StyleGAN3.

Note: some details will not be mentioned since I want to make it short and only talk about the architectural changes and their purposes.

Table of Contents

• StyleGAN
  ∘ Step 1: Mapping and Styles
  ∘ Step 2: Constant Input
  ∘ Step 3: Noise Inputs
  ∘ Step 4: Style Mixing
• StyleGAN2
  ∘ Reason
  ∘ Original StyleGAN Design
  ∘ Changes #1
  ∘ Changes #2
  ∘ Changes #3
  ∘ StyleGAN2 Results
• StyleGAN2-ADA
  ∘ Stochastic Discriminator Augmentation
  ∘ Adaptive Discriminator Augmentation
  ∘ StyleGAN2-ADA Results
• StyleGAN3
  ∘ Texture Sticking
  ∘ Reason: Positional References
  ∘ Goal
  ∘ Redesigning Network Layers
  ∘ Changes
  ∘ StyleGAN3 Results
• Summary
• References

StyleGAN

The purpose of StyleGAN is to synthesize photorealistic/high-fidelity images.

StyleGAN generator architecture

The architecture of the StyleGAN generator might look complicated at the first…