Recognition: unknown
LayoutGAN: Generating Graphic Layouts with Wireframe Discriminators
read the original abstract
Layout is important for graphic design and scene generation. We propose a novel Generative Adversarial Network, called LayoutGAN, that synthesizes layouts by modeling geometric relations of different types of 2D elements. The generator of LayoutGAN takes as input a set of randomly-placed 2D graphic elements and uses self-attention modules to refine their labels and geometric parameters jointly to produce a realistic layout. Accurate alignment is critical for good layouts. We thus propose a novel differentiable wireframe rendering layer that maps the generated layout to a wireframe image, upon which a CNN-based discriminator is used to optimize the layouts in image space. We validate the effectiveness of LayoutGAN in various experiments including MNIST digit generation, document layout generation, clipart abstract scene generation and tangram graphic design.
This paper has not been read by Pith yet.
Forward citations
Cited by 3 Pith papers
-
Evolving Thematic Map Design in Academic Cartography: A Thirty-Year Study Based on Multilingual Journals
Thematic maps in academic Chinese- and English-language journals have converged on similar design conventions over 30 years, featuring muted colors, high main-map occupancy, and increasing element and hue diversity wh...
-
GAN-based Domain Adaptation for Image-aware Layout Generation in Advertising Poster Design
PDA-GAN with pixel discriminator bridges domain gap from inpainted posters to generate SOTA image-aware layouts on a new 60k-pair CGL-Dataset.
-
CPT: Controllable and Editable Design Variations with Language Models
CPT is a fine-tuned language model that uses Creative Markup Language representations of professional designs to generate controllable, stylistically coherent, and fully editable design variations.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.