Edge Prediction for Roof Wireframe Reconstruction with Transformers
Gustav Hanning, Ludvig Dillén, Jonathan Astermark, Johanna Lidholm, Viktor LarssonCentre for Mathematical Sciences, Lund University {gustav.hanning,ludvig.dillen,jonathan.astermark,johanna.lidholm,viktor.larsson}@math.lth.se · 2026
Read Edge Prediction for Roof Wireframe Reconstruction with Transformers by Gustav Hanning, Ludvig Dillén, Jonathan Astermark, Johanna Lidholm, Viktor LarssonCentre for Mathematical Sciences, Lund University {gustav.hanning,ludvig.dillen,jonathan.astermark,johanna.lidholm,viktor.larsson}@math.lth.se (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2606-02406.txt.
Description
Hanning and co-authors rebuild house roofs as 3D wireframes from sparse data. Their entry targets the S23DR Challenge 2026. Inputs are sparse point clouds from structure-from-motion. Ground-level semantic labels and depth maps join them. The method is a Transformer encoder-decoder in the DETR style.
The point cloud is thinned by semantic priority before use. Gestalt and ADE20k class features enrich each point. Points also draw features from a frozen autoencoder through projection. Learned queries then decode straight into 3D wireframe edges. Cross-attention drives that decoding step. On the HoHo 22k dataset, the method beats crafted and learned baselines. Its Hybrid Structure Score reaches 0.6476. That score took second place on the private leaderboard.
Catalog record
| CREATOR | Gustav Hanning, Ludvig Dillén, Jonathan Astermark, Johanna Lidholm, Viktor LarssonCentre for Mathematical Sciences, Lund University {gustav.hanning,ludvig.dillen,jonathan.astermark,johanna.lidholm,viktor.larsson}@math.lth.se |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 28 KB |
| LICENSE | http://creativecommons.org/licenses/by-sa/4.0/ |
| RIGHTS | CC BY SA |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2606.02406