Analytical Interpretation of Latent Codes in InfoGAN with SAR Images

TL;DR

This paper provides an analytical nonlinear model explaining how latent codes in InfoGAN influence SAR image properties, enabling precise property manipulation through latent code estimation.

Contribution

It introduces a novel analytical nonlinear model that clarifies the causal relationship between latent codes and SAR image properties in InfoGAN.

Findings

Latent codes affect SAR image properties in a nonlinear, disentangled manner.

Properties can be predicted from latent codes and vice versa.

The model enables targeted property editing in SAR image synthesis.

Abstract

Generative Adversarial Networks (GANs) can synthesize abundant photo-realistic synthetic aperture radar (SAR) images. Some recent GANs (e.g., InfoGAN), are even able to edit specific properties of the synthesized images by introducing latent codes. It is crucial for SAR image synthesis since the targets in real SAR images are with different properties due to the imaging mechanism. Despite the success of InfoGAN in manipulating properties, there still lacks a clear explanation of how these latent codes affect synthesized properties, thus editing specific properties usually relies on empirical trials, unreliable and time-consuming. In this paper, we show that latent codes are disentangled to affect the properties of SAR images in a non-linear manner. By introducing some property estimators for latent codes, we are able to provide a completely analytical nonlinear model to decompose the entangled causality between latent codes and different properties. The qualitative and quantitative experimental results further reveal that the properties can be calculated by latent codes, inversely, the satisfying latent codes can be estimated given desired properties. In this case, properties can be manipulated by latent codes as we expect.