Vector Prism: Animating Vector Graphics by Stratifying Semantic Structure
This work solves the challenge of reliable SVG animation for web design, enabling more dynamic and interpretable interactions with vision-language models, though it is incremental in improving existing VLM capabilities.
The paper tackles the problem of automating SVG animation by addressing the fragmentation of vector graphics into low-level shapes, which hinders vision-language models. It introduces a framework that recovers semantic structure through statistical aggregation of weak predictions, resulting in substantial gains in animation coherence over existing approaches.
Scalable Vector Graphics (SVG) are central to modern web design, and the demand to animate them continues to grow as web environments become increasingly dynamic. Yet automating the animation of vector graphics remains challenging for vision-language models (VLMs) despite recent progress in code generation and motion planning. VLMs routinely mis-handle SVGs, since visually coherent parts are often fragmented into low-level shapes that offer little guidance of which elements should move together. In this paper, we introduce a framework that recovers the semantic structure required for reliable SVG animation and reveals the missing layer that current VLM systems overlook. This is achieved through a statistical aggregation of multiple weak part predictions, allowing the system to stably infer semantics from noisy predictions. By reorganizing SVGs into semantic groups, our approach enables VLMs to produce animations with far greater coherence. Our experiments demonstrate substantial gains over existing approaches, suggesting that semantic recovery is the key step that unlocks robust SVG animation and supports more interpretable interactions between VLMs and vector graphics.