Towards Aligned Layout Generation via Diffusion Model with Aesthetic Constraints

TL;DR

This paper introduces LACE, a continuous diffusion model that generates visually appealing graphic layouts by incorporating aesthetic constraints, outperforming previous methods in quality and flexibility.

Contribution

LACE is a unified continuous diffusion model that handles diverse layout generation tasks with aesthetic constraints, advancing controllability and quality in graphic design layout synthesis.

Findings

LACE outperforms existing state-of-the-art models in layout quality.

The continuous diffusion approach enables incorporation of aesthetic constraints.

LACE effectively handles various layout generation tasks.

Abstract

Controllable layout generation refers to the process of creating a plausible visual arrangement of elements within a graphic design (e.g., document and web designs) with constraints representing design intentions. Although recent diffusion-based models have achieved state-of-the-art FID scores, they tend to exhibit more pronounced misalignment compared to earlier transformer-based models. In this work, we propose the $\textbf{LA}$yout $\textbf{C}$onstraint diffusion mod$\textbf{E}$l (LACE), a unified model to handle a broad range of layout generation tasks, such as arranging elements with specified attributes and refining or completing a coarse layout design. The model is based on continuous diffusion models. Compared with existing methods that use discrete diffusion models, continuous state-space design can enable the incorporation of differentiable aesthetic constraint functions in training. For conditional generation, we introduce conditions via masked input. Extensive experiment results show that LACE produces high-quality layouts and outperforms existing state-of-the-art baselines.