Unit Region Encoding: A Unified and Compact Geometry-aware Representation for Floorplan Applications

2025-01-19

AI Summary

To address the challenge of jointly modeling shape and positional information in architectural floorplan representation—while suffering from low spatial efficiency—this paper proposes Unit Region Encoding (URE), a geometrically aware, compact, and unified floorplan representation method. URE introduces boundary-adaptive unit partitioning and geometry-aware density map–driven clustering, enabling the first density-guided adaptive region slicing, which replaces conventional over-segmented raster grids and coarse-grained graph structures. The framework comprises URE-Net, an end-to-end encoding network, and a multi-task joint representation learning scheme. Extensive experiments demonstrate state-of-the-art performance across floorplan understanding, generation, and metric learning tasks. Ablation studies validate the effectiveness of the adaptive slicing strategy. Moreover, URE yields highly compact encodings, supports cross-task transferability, and enables high-fidelity reconstruction, exhibiting strong scalability and generalization.

Technology Category

• Reasoning under Uncertainty: Uncertainty Representations
• Computer Vision: Representation Learning for Vision
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Search and Retrieval-Augmented AI: Web query analysis, representation and understanding
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

We present the Unit Region Encoding of floorplans, which is a unified and compact geometry-aware encoding representation for various applications, ranging from interior space planning, floorplan metric learning to floorplan generation tasks. The floorplans are represented as the latent encodings on a set of boundary-adaptive unit region partition based on the clustering of the proposed geometry-aware density map. The latent encodings are extracted by a trained network (URE-Net) from the input dense density map and other available semantic maps. Compared to the over-segmented rasterized images and the room-level graph structures, our representation can be flexibly adapted to different applications with the sliced unit regions while achieving higher accuracy performance and better visual quality. We conduct a variety of experiments and compare to the state-of-the-art methods on the aforementioned applications to validate the superiority of our representation, as well as extensive ablation studies to demonstrate the effect of our slicing choices.

Problem

• Floor Design Optimization
• Shape and Position Representation
• Indoor Space Understanding

Innovation

• Unit Region Encoding
• Geometric Perception Density Map
• URE-Net Network