StyLandGAN: A StyleGAN based Landscape Image Synthesis using Depth-map

TL;DR

StyLandGAN introduces a novel landscape image synthesis framework that utilizes depth maps for higher expressive power, employing a 2-phase inference pipeline to enhance diversity and user control, outperforming existing methods.

Contribution

The paper presents StyLandGAN, a new StyleGAN-based model that synthesizes landscape images from depth maps and introduces a 2-phase inference process for improved diversity and user control.

Findings

Outperforms existing methods in image quality and diversity.

Effectively reflects user intentions through local part shifting.

Achieves higher depth-accuracy in synthesized images.

Abstract

Despite recent success in conditional image synthesis, prevalent input conditions such as semantics and edges are not clear enough to express `Linear (Ridges)' and `Planar (Scale)' representations. To address this problem, we propose a novel framework StyLandGAN, which synthesizes desired landscape images using a depth map which has higher expressive power. Our StyleLandGAN is extended from the unconditional generation model to accept input conditions. We also propose a '2-phase inference' pipeline which generates diverse depth maps and shifts local parts so that it can easily reflect user's intend. As a comparison, we modified the existing semantic image synthesis models to accept a depth map as well. Experimental results show that our method is superior to existing methods in quality, diversity, and depth-accuracy.