Unsupervised Domain Adaption with Pixel-level Discriminator for Image-aware Layout Generation

TL;DR

This paper introduces PDA-GAN, a novel unsupervised domain adaptation method with a pixel-level discriminator, to generate high-quality, image-aware graphic layouts for advertising posters, addressing domain gaps in training data.

Contribution

The paper proposes a new GAN model with a pixel-level discriminator for unsupervised domain adaptation in layout generation, improving quality and realism.

Findings

Achieves state-of-the-art performance in layout generation

Generates high-quality, image-aware graphic layouts

Effectively bridges domain gap in training data

Abstract

Layout is essential for graphic design and poster generation. Recently, applying deep learning models to generate layouts has attracted increasing attention. This paper focuses on using the GAN-based model conditioned on image contents to generate advertising poster graphic layouts, which requires an advertising poster layout dataset with paired product images and graphic layouts. However, the paired images and layouts in the existing dataset are collected by inpainting and annotating posters, respectively. There exists a domain gap between inpainted posters (source domain data) and clean product images (target domain data). Therefore, this paper combines unsupervised domain adaption techniques to design a GAN with a novel pixel-level discriminator (PD), called PDA-GAN, to generate graphic layouts according to image contents. The PD is connected to the shallow level feature map and computes the GAN loss for each input-image pixel. Both quantitative and qualitative evaluations demonstrate that PDA-GAN can achieve state-of-the-art performances and generate high-quality image-aware graphic layouts for advertising posters.