Jain et al. (2022) | VectorFusion: Text-to-SVG by Abstracting Pixel-Based Diffusion Models¶

Jain et al. propose a text-conditioned diffusion model trained on pixel representations of images which can be used to generate SVG-exportable vector graphics. Their work is based on the differentiable render DiffVG by Li et al. (2020)

Available resources at a glance

arXiv URL to the paper
Github repository with code: not available (yet)
Project page: https://ajayj.com/vectorfusion

Fig. 58 Screenshot of the Jain et al. (2022) paper¶

Deep Learning with Vector Graphics

Jain et al. (2022) | VectorFusion: Text-to-SVG by Abstracting Pixel-Based Diffusion Models¶