MagicGeo: Training-Free Text-Guided Geometric Diagram Generation

2025-02-19

AI Summary

This work addresses the high annotation cost of manual geometric diagram drawing and the low geometric fidelity of existing AIGC models in text-to-geometry generation. We propose the first training-free, text-driven geometric diagram generation framework. Methodologically, it synergistically integrates the zero-shot semantic parsing capability of large language models (LLMs) with formal geometric solvers (e.g., Z3, symbolic algebra) to perform coordinate-wise co-optimization—ensuring both semantic accuracy and strict satisfaction of geometric constraints. Our contributions are threefold: (1) introducing MagicGeoBench, the first benchmark dedicated to geometric diagram generation; (2) enabling multi-step constructive diagram synthesis and supporting real-time interactive editing; and (3) achieving a 37% absolute improvement in geometric correctness over state-of-the-art methods on MagicGeoBench. This advances structured visual generation research, particularly for educational applications and interpretable AI.

Technology Category

• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning
• Computer Vision: Visual Reasoning & Symbolic Representations
• Natural Language Processing: Sentence-level Semantics, Textual Inference, etc.

Application Category

• 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
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information

Abstract

Geometric diagrams are critical in conveying mathematical and scientific concepts, yet traditional diagram generation methods are often manual and resource-intensive. While text-to-image generation has made strides in photorealistic imagery, creating accurate geometric diagrams remains a challenge due to the need for precise spatial relationships and the scarcity of geometry-specific datasets. This paper presents MagicGeo, a training-free framework for generating geometric diagrams from textual descriptions. MagicGeo formulates the diagram generation process as a coordinate optimization problem, ensuring geometric correctness through a formal language solver, and then employs coordinate-aware generation. The framework leverages the strong language translation capability of large language models, while formal mathematical solving ensures geometric correctness. We further introduce MagicGeoBench, a benchmark dataset of 220 geometric diagram descriptions, and demonstrate that MagicGeo outperforms current methods in both qualitative and quantitative evaluations. This work provides a scalable, accurate solution for automated diagram generation, with significant implications for educational and academic applications.

Problem

• Automated geometric diagram generation
• Training-free text-guided framework
• Ensuring geometric correctness via formal language solver

Innovation

• Training-free geometric diagram generation
• Coordinate optimization for spatial accuracy
• Leveraging large language models