Enhancing Polygonal Building Segmentation via Oriented Corners

TL;DR

This paper presents OriCornerNet, a novel deep neural network that directly extracts accurate, regular, and simplified building polygons from overhead images by predicting masks, corners, and orientations, reducing post-processing needs.

Contribution

Introduces OriCornerNet, a deep model that directly predicts building polygons using oriented corners and iterative refinement, improving accuracy and regularity over existing methods.

Findings

Outperforms state-of-the-art in building segmentation accuracy.

Produces more regular and simplified building polygons.

Effective on SpaceNet Vegas and CrowdAI-small datasets.

Abstract

The growing demand for high-resolution maps across various applications has underscored the necessity of accurately segmenting building vectors from overhead imagery. However, current deep neural networks often produce raster data outputs, leading to the need for extensive post-processing that compromises the fidelity, regularity, and simplicity of building representations. In response, this paper introduces a novel deep convolutional neural network named OriCornerNet, which directly extracts delineated building polygons from input images. Specifically, our approach involves a deep model that predicts building footprint masks, corners, and orientation vectors that indicate directions toward adjacent corners. These predictions are then used to reconstruct an initial polygon, followed by iterative refinement using a graph convolutional network that leverages semantic and geometric features. Our method inherently generates simplified polygons by initializing the refinement process with predicted corners. Also, including geometric information from oriented corners contributes to producing more regular and accurate results. Performance evaluations conducted on SpaceNet Vegas and CrowdAI-small datasets demonstrate the competitive efficacy of our approach compared to the state-of-the-art in building segmentation from overhead imagery.