FGSSNet: Feature-Guided Semantic Segmentation of Real World Floorplans

2025-07-14

AI Summary

To address the limited generalization capability of wall segmentation in real-world floor plans, this paper proposes a feature-guided semantic segmentation method. The core innovation is a multi-head feature extractor that jointly predicts wall texture and width; the compressed domain-specific feature maps are then injected into the U-Net latent space to explicitly model complex wall structures—including thin, discontinuous, and intersecting walls. This feature-guidance mechanism is end-to-end optimized with the U-Net backbone, requiring no additional annotations. Experiments on multiple real-world floor plan datasets demonstrate that our method significantly outperforms standard U-Net, achieving consistent improvements in IoU (+3.2–5.7%) and boundary accuracy. Moreover, it exhibits superior robustness and generalization across diverse drawing styles and low-quality plans.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Feature Construction/Reformulation
• Planning, Routing, and Scheduling: Planning with Language Models

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Web Mining and Content Analysis: Large pretrained models with web data
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

We introduce FGSSNet, a novel multi-headed feature-guided semantic segmentation (FGSS) architecture designed to improve the generalization ability of wall segmentation on floorplans. FGSSNet features a U-Net segmentation backbone with a multi-headed dedicated feature extractor used to extract domain-specific feature maps which are injected into the latent space of U-Net to guide the segmentation process. This dedicated feature extractor is trained as an encoder-decoder with selected wall patches, representative of the walls present in the input floorplan, to produce a compressed latent representation of wall patches while jointly trained to predict the wall width. In doing so, we expect that the feature extractor encodes texture and width features of wall patches that are useful to guide the wall segmentation process. Our experiments show increased performance by the use of such injected features in comparison to the vanilla U-Net, highlighting the validity of the proposed approach.

Problem

• Improving wall segmentation generalization in floorplans
• Injecting domain-specific features into U-Net latent space
• Jointly training feature extractor for texture and width prediction

Innovation

• Multi-headed feature-guided segmentation architecture
• U-Net backbone with domain-specific feature injection
• Encoder-decoder trained on wall patches