B'ezier Splatting for Fast and Differentiable Vector Graphics

2025-03-20

AI Summary

Existing differentiable vector graphics (VG) representations suffer from high computational overhead and degraded image quality during high-resolution rendering. This paper proposes Bézier Rasterization: a novel differentiable rasterizer that models shapes using Bézier curves, employs 2D Gaussian splatting for differentiable rasterization, and—crucially—introduces boundary position gradients to enhance geometric fidelity. To mitigate optimization stagnation in local minima, we design an adaptive curve pruning and densification mechanism. Compared to DiffVG, our method achieves 20× and 150× speedups in forward and backward rasterization, respectively, and delivers a 10× end-to-end optimization acceleration. Consequently, it significantly improves both visual quality and convergence efficiency for high-resolution vector synthesis.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Search and Optimization: Sampling/Simulation-based Search
• Machine Learning: Optimization

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Responsible Web: Data and user privacy-enhancing technologies for the Web
• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines

Abstract

Differentiable vector graphics (VGs) are widely used in image vectorization and vector synthesis, while existing representations are costly to optimize and struggle to achieve high-quality rendering results for high-resolution images. This work introduces a new differentiable VG representation, dubbed B'ezier splatting, that enables fast yet high-fidelity VG rasterization. B'ezier splatting samples 2D Gaussians along B'ezier curves, which naturally provide positional gradients at object boundaries. Thanks to the efficient splatting-based differentiable rasterizer, B'ezier splatting achieves over 20x and 150x faster per forward and backward rasterization step for open curves compared to DiffVG. Additionally, we introduce an adaptive pruning and densification strategy that dynamically adjusts the spatial distribution of curves to escape local minima, further improving VG quality. Experimental results show that B'ezier splatting significantly outperforms existing methods with better visual fidelity and 10x faster optimization speed.

Problem

• Improves optimization speed of differentiable vector graphics
• Enhances rendering quality for high-resolution images
• Introduces adaptive pruning for better curve distribution

Innovation

• Bézier splatting for fast VG rasterization
• 2D Gaussians sampled along Bézier curves
• Adaptive pruning and densification strategy