ACPV-Net: All-Class Polygonal Vectorization for Seamless Vector Map Generation from Aerial Imagery

CVPR 2026
Weiqin Jiao, Hao Cheng, George Vosselman, Claudio Persello
Faculty of Geo-Information Science and Earth Observation (ITC), University of Twente
Contact: w.jiao@utwente.nl
arXiv Code Dataset Supplement
ACPV-Net teaser figure

To generate an all-class vector basemap, existing single-class polygonization methods require per-class inference and stitching, resulting in gaps and overlaps across classes. ACPV-Net is the first fully automatic framework that produces a seamless basemap with shared-edge consistency and avoids gaps and overlaps in a single run.

Abstract

We tackle the problem of generating a complete vector map representation from aerial imagery in a single run: producing polygons for all land-cover classes with shared boundaries and without gaps or overlaps. Existing polygonization methods are typically class-specific; extending them to multiple classes via per-class runs commonly leads to topological inconsistencies, such as duplicated edges, gaps, and overlaps. We formalize this new task as All-Class Polygonal Vectorization (ACPV) and release the first public benchmark, Deventer-512, with standardized metrics jointly evaluating semantic fidelity, geometric accuracy, vertex efficiency, per-class topological fidelity and global topological consistency. To realize ACPV, we propose ACPV-Net, a unified framework introducing a novel Semantically Supervised Conditioning (SSC) mechanism coupling semantic perception with geometric primitive generation, along with a topological reconstruction that enforces shared-edge consistency by design. While enforcing such strict topological constraints, ACPV-Net surpasses all class-specific baselines in polygon quality across classes on Deventer-512. It also applies to single-class polygonal vectorization without any architectural modification, achieving the best-reported results on WHU-Building.

Contributions

  • We address the vector basemap generation problem by formulating it as All-Class Polygonal Vectorization (ACPV): generating a single, globally consistent planar partition of the image domain into per-class polygons that satisfy strict topological consistency.
  • We introduce ACPV-Net, the first fully automatic framework that, in a single run, produces topologically consistent vector basemaps from aerial imagery.
  • We release Deventer-512, the first public ACPV benchmark with high-resolution imagery, fine-grained polygons across five land-cover classes, a standardized evaluation protocol, and subsets for challenging conditions.
  • ACPV-Net achieves gap-/overlap-free topology and significantly outperforms strong class-specific baselines across land-cover classes on Deventer-512. It also applies, without any architectural change, to single-class polygonal vectorization on the WHU-Building dataset while maintaining excellent performance.

Method

ACPV-Net unifies semantically supervised conditioning and proposition-driven topological reconstruction: the former produces coherent semantic-geometric evidence through diffusion-based vertex generation under semantic supervision, the latter deterministically reconstructs a topology-consistent vector basemap via overdense PSLG construction and vertex-guided subset selection.

ACPV-Net network architecture

Qualitative Comparison

A challenging multi-class scene with complex geometry. ACPV-Net preserves class boundaries while maintaining seamless topology.

Input

Geometric complexity input

GT

Geometric complexity ground truth

GCP

Geometric complexity GCP

HiSup

Geometric complexity HiSup

TopDiG

Geometric complexity TopDiG

ACPV-Net (Ours)

Geometric complexity ACPV-Net

Thin and elongated structures are sensitive to topological breaks and vertex redundancy. ACPV-Net recovers these details while keeping polygons coherent.

Input

Thin structures input

GT

Thin structures ground truth

GCP

Thin structures GCP

HiSup

Thin structures HiSup

TopDiG

Thin structures TopDiG

ACPV-Net (Ours)

Thin structures ACPV-Net

Under difficult appearance conditions, ACPV-Net still infers more consistent polygonal maps and shared boundaries than existing baselines.

Input

Weak visual condition input

GT

Weak visual condition ground truth

GCP

Weak visual condition GCP

HiSup

Weak visual condition HiSup

TopDiG

Weak visual condition TopDiG

ACPV-Net (Ours)

Weak visual condition ACPV-Net

Acknowledgements

This publication is part of the project “Learning from old maps to create new ones”, with project number 19206 of the Open Technology Programme, which is financed by the Dutch Research Council (NWO), The Netherlands. We thank Prof. Sander Oude Elberink, Geethanjali Anjanappa, Dr. Yaping Lin, and Dr. Abhisek Maiti for their support within the project, especially in data preparation and related research activities.

BibTeX

@misc{jiao2026acpvnetallclasspolygonalvectorization,
  title={ACPV-Net: All-Class Polygonal Vectorization for Seamless Vector Map Generation from Aerial Imagery},
  author={Weiqin Jiao and Hao Cheng and George Vosselman and Claudio Persello},
  year={2026},
  eprint={2603.16616},
  archivePrefix={arXiv},
  primaryClass={cs.CV},
  url={https://arxiv.org/abs/2603.16616},
}

@inproceedings{jiao2026acpvnet,
  title={ACPV-Net: All-Class Polygonal Vectorization for Seamless Vector Map Generation from Aerial Imagery},
  author={Jiao, Weiqin and Cheng, Hao and Vosselman, George and Persello, Claudio},
  booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  year={2026},
  note={Accepted, camera-ready version pending}
}