CVGRLGNov 21, 2019

Semantic Hierarchy Emerges in Deep Generative Representations for Scene Synthesis

arXiv:1911.09267v3212 citations
Originality Incremental advance
AI Analysis

This provides insights into how GANs learn and organize information, which is incremental for improving interpretability and control in image synthesis.

The study demonstrates that deep generative models like StyleGAN and BigGAN develop a structured semantic hierarchy in their representations, where early layers influence spatial layout, middle layers control objects, and later layers handle attributes and colors, enabling semantic scene manipulation.

Despite the success of Generative Adversarial Networks (GANs) in image synthesis, there lacks enough understanding on what generative models have learned inside the deep generative representations and how photo-realistic images are able to be composed of the layer-wise stochasticity introduced in recent GANs. In this work, we show that highly-structured semantic hierarchy emerges as variation factors from synthesizing scenes from the generative representations in state-of-the-art GAN models, like StyleGAN and BigGAN. By probing the layer-wise representations with a broad set of semantics at different abstraction levels, we are able to quantify the causality between the activations and semantics occurring in the output image. Such a quantification identifies the human-understandable variation factors learned by GANs to compose scenes. The qualitative and quantitative results further suggest that the generative representations learned by the GANs with layer-wise latent codes are specialized to synthesize different hierarchical semantics: the early layers tend to determine the spatial layout and configuration, the middle layers control the categorical objects, and the later layers finally render the scene attributes as well as color scheme. Identifying such a set of manipulatable latent variation factors facilitates semantic scene manipulation.

Code Implementations2 repos
Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes