Research map / Video / Video generation
VAnim: Rendering-Aware Sparse State Modeling for Structure-Preserving Vector Animation
Guotao Liang, Zhangcheng Wang, Chuang Wang, Juncheng Hu, Haitao Zhou, Junhua Liu, Jing Zhang, Dong Xu, Qian Yu
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add rendering-aware GRPO after SFT (ablation) | VAnim · SVGAnim-Test | Semantic Alignment: 0.268 -> 0.281 | Improved |
| add rendering-aware GRPO after SFT (ablation) | VAnim · SVGAnim-Test | Success Rate: 95.2% -> 100.0% | Improved |
| add Structure-Bound CoT planning (ablation) | VAnim · SVGAnim-Test | Semantic Alignment: 0.255 -> 0.281 | Improved |
| add Structure-Bound CoT planning (ablation) | VAnim · SVGAnim-Test | Success Rate: 98.6% -> 100.0% | Improved |
| use Sparse State Updates instead of frame-by-frame generation (ablation) | VAnim · SVGAnim-Test | Success Rate: 62.3% -> 100.0% | Improved |
| use Sparse State Updates instead of full SVG per frame | VAnim · 24-frame animation | token count: 86.0k -> 9.2k | Improved |
| use VAnim over GPT-5.2 and Gemini 3 Pro | VAnim · SVGAnim-Test | Semantic Alignment: 0.243 -> 0.281 (vs Gemini 3 Pro) | Improved |
| remove input image (ablation) | VAnim · SVGAnim-Test | SSIM and temporal smoothness | Hurt |
| vary GRPO group size 4/8/16 (ablation) | VAnim · SVGAnim-Test | semantic alignment vs identity preservation | Mixed |
Closest papers by meaning
- Rendering-Aware Reinforcement Learning for Vector Graphics Generation (NeurIPS 2025 · similarity 0.89)
- Animation2Code: Evaluating Temporal Visual Reasoning in Video-to-Code Generation (NeurIPS 2026 · similarity 0.86)
- Chat2SVG: Vector Graphics Generation with Large Language Models and Image Diffusion Models (CVPR 2025 · similarity 0.86)
- DrawingsDreamer: A Unified Multi-View Engineering Drawings Generation Model (NeurIPS 2026 · similarity 0.85)
- Breaking the SFT Plateau: Multimodal Structured Reinforcement Learning for Chart-to-Code Generation (ICLR 2026 · similarity 0.85)
- VFIG: Vectorizing Complex Figures in SVG with Vision-Language Models (NeurIPS 2026 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
