How to Correct Stroke Rendering Using GPU-Parallel Algorithms

This paper introduces a fully parallel GPU algorithm for stroke expansion—the process of generating outlines for stroked vector paths. The method leverages Euler spiral approximations, compact binary path encoding, and error-bounded curve metrics to achieve strongly correct stroke rendering while minimizing preprocessing. By offloading computation to GPU compute shaders, it dramatically improves the speed and accuracy of vector stroke rendering compared to CPU-based methods, enabling real-time performance for complex graphics scenes.