Semantic Document Derendering: SVG Reconstruction via Vision-Language Modeling

2025-11-17

AI Summary

Converting static raster documents (e.g., slides) into structured vector formats (SVG) often results in loss of high-level semantic structure and ambiguous separation of textual and graphical elements. Method: This paper proposes SliDer, the first framework to employ vision-language models (VLMs) for semantic-level document reverse rendering. It leverages multimodal understanding to accurately detect document elements, parse their semantic and layout attributes, and iteratively generate editable SVG code. To support this task, we introduce Slide2SVG—the first annotated dataset specifically designed for slide-to-SVG conversion—and evaluate reconstruction quality using perceptual metrics such as LPIPS. Contribution/Results: Experiments demonstrate that SliDer outperforms the strongest zero-shot VLM baseline on 82.9% of test samples, achieving an LPIPS score of 0.069. The method significantly improves structural fidelity, visual quality, and post-conversion editability of generated SVGs.

Technology Category

• Computer Vision: Language and Vision
• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: Summarization

Application Category

• Search and Retrieval-Augmented AI: Large language models for search
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Multimedia documents such as slide presentations and posters are designed to be interactive and easy to modify. Yet, they are often distributed in a static raster format, which limits editing and customization. Restoring their editability requires converting these raster images back into structured vector formats. However, existing geometric raster-vectorization methods, which rely on low-level primitives like curves and polygons, fall short at this task. Specifically, when applied to complex documents like slides, they fail to preserve the high-level structure, resulting in a flat collection of shapes where the semantic distinction between image and text elements is lost. To overcome this limitation, we address the problem of semantic document derendering by introducing SliDer, a novel framework that uses Vision-Language Models (VLMs) to derender slide images as compact and editable Scalable Vector Graphic (SVG) representations. SliDer detects and extracts attributes from individual image and text elements in a raster input and organizes them into a coherent SVG format. Crucially, the model iteratively refines its predictions during inference in a process analogous to human design, generating SVG code that more faithfully reconstructs the original raster upon rendering. Furthermore, we introduce Slide2SVG, a novel dataset comprising raster-SVG pairs of slide documents curated from real-world scientific presentations, to facilitate future research in this domain. Our results demonstrate that SliDer achieves a reconstruction LPIPS of 0.069 and is favored by human evaluators in 82.9% of cases compared to the strongest zero-shot VLM baseline.

Problem

• Converting static raster slide images into editable vector SVG formats
• Preserving semantic structure between image and text elements during conversion
• Overcoming limitations of geometric methods that lose high-level document structure

Innovation

• Uses Vision-Language Models for SVG reconstruction
• Iteratively refines predictions during inference process
• Detects and organizes elements into coherent SVG format