Spaces:
Running on Zero
Running on Zero
| title: Pixels Points Polygons Building Vectorizer | |
| emoji: 🏙️ | |
| colorFrom: purple | |
| colorTo: blue | |
| sdk: gradio | |
| sdk_version: 6.20.0 | |
| app_file: app.py | |
| short_description: Extract building polygons from aerial images (P3 / Pix2Poly) | |
| python_version: "3.12" | |
| startup_duration_timeout: 1h | |
| pinned: false | |
| # Pixels, Points & Polygons — Building Vectorizer | |
| Interactive demo of the **Pix2Poly image model** from | |
| [*The P³ Dataset: Pixels, Points and Polygons for Multimodal Building Vectorization*](https://huggingface.co/papers/2505.15379) | |
| (Sulzer, Duan, Girard & Lafarge, 2025). | |
| Upload a nadir aerial RGB tile and the model predicts closed **building outline | |
| polygons**, drawn as a vector overlay. The pipeline is a DINO ViT-S/8 encoder → | |
| transformer polygon decoder (autoregressive vertex generation) → optimal-transport | |
| permutation head that connects vertices into polygons. | |
| - **Weights:** [`rsi/PixelsPointsPolygons`](https://huggingface.co/rsi/PixelsPointsPolygons) (checkpoint `v4_image_vit_bs4x16`) | |
| - **Code:** [github.com/raphaelsulzer/pixelspointspolygons](https://github.com/raphaelsulzer/pixelspointspolygons) | |
| Inputs are resized to 224×224; the model was trained on 25 cm ground-sampling-distance | |
| imagery, so real-world aerial tiles at similar scale work best. Only the image modality | |
| is served here (the LiDAR / fusion variants need Open3D-ML and custom CUDA ops). | |