GSPlane: Concise and Accurate Planar Reconstruction via Structured Representation

2025-10-19

AI Summary

Existing Gaussian Splatting (GS) methods suffer from low geometric accuracy, non-smooth surfaces, and poor mesh connectivity in planar region reconstruction. To address these issues, we propose a structured planar reconstruction framework: first, leveraging segmentation and normal prediction networks to extract planar priors; second, designing structured planar Gaussian primitives with explicit geometric consistency constraints and a dynamic Gaussian reclassification mechanism to model planar topology during optimization; and third, refining mesh layout to enhance structural quality. Our method preserves rendering fidelity while significantly improving planar geometric accuracy, reducing vertex and face counts, and enabling disentangled, editable structured planar representations. This establishes a novel paradigm for semantic 3D scene reconstruction and editing.

Technology Category

• Planning, Routing, and Scheduling: Replanning and Plan Repair
• Machine Learning: Structured Learning
• Computer Vision: Segmentation

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Semantics and Knowledge: Scalable techniques for the creation, curation, publication, maintenance, and consumption of large, Web-based, structured, reusable, knowledge graphs and ontologies
• Search and Retrieval-Augmented AI: Web query analysis, representation and understanding

Abstract

Planes are fundamental primitives of 3D sences, especially in man-made environments such as indoor spaces and urban streets. Representing these planes in a structured and parameterized format facilitates scene editing and physical simulations in downstream applications. Recently, Gaussian Splatting (GS) has demonstrated remarkable effectiveness in the Novel View Synthesis task, with extensions showing great potential in accurate surface reconstruction. However, even state-of-the-art GS representations often struggle to reconstruct planar regions with sufficient smoothness and precision. To address this issue, we propose GSPlane, which recovers accurate geometry and produces clean and well-structured mesh connectivity for plane regions in the reconstructed scene. By leveraging off-the-shelf segmentation and normal prediction models, GSPlane extracts robust planar priors to establish structured representations for planar Gaussian coordinates, which help guide the training process by enforcing geometric consistency. To further enhance training robustness, a Dynamic Gaussian Re-classifier is introduced to adaptively reclassify planar Gaussians with persistently high gradients as non-planar, ensuring more reliable optimization. Furthermore, we utilize the optimized planar priors to refine the mesh layouts, significantly improving topological structure while reducing the number of vertices and faces. We also explore applications of the structured planar representation, which enable decoupling and flexible manipulation of objects on supportive planes. Extensive experiments demonstrate that, with no sacrifice in rendering quality, the introduction of planar priors significantly improves the geometric accuracy of the extracted meshes across various baselines.

Problem

• Reconstructs planar regions with smoothness and precision in 3D scenes.
• Establishes structured planar representations using geometric consistency priors.
• Improves mesh topological structure and reduces vertex-face count.

Innovation

• Leveraging planar priors to guide Gaussian training process
• Introducing dynamic Gaussian reclassifier for robust optimization
• Refining mesh layouts using optimized planar priors