Diffusion-based Document Layout Generation
This work addresses the problem of generating realistic document layouts for design and automation applications, representing an incremental improvement with a novel metric.
The paper tackles document layout generation by developing a diffusion-based approach that operates in the sequence domain to create complex layouts, and introduces a new metric called Document Earth Mover's Distance (Doc-EMD) to better evaluate heterogeneous document designs. The results show that the method is comparable to or outperforms previous approaches across various datasets, and the metric improves differentiation in specific cases.
We develop a diffusion-based approach for various document layout sequence generation. Layout sequences specify the contents of a document design in an explicit format. Our novel diffusion-based approach works in the sequence domain rather than the image domain in order to permit more complex and realistic layouts. We also introduce a new metric, Document Earth Mover's Distance (Doc-EMD). By considering similarity between heterogeneous categories document designs, we handle the shortcomings of prior document metrics that only evaluate the same category of layouts. Our empirical analysis shows that our diffusion-based approach is comparable to or outperforming other previous methods for layout generation across various document datasets. Moreover, our metric is capable of differentiating documents better than previous metrics for specific cases.