rsi commited on
Commit
9cf9ecf
·
verified ·
1 Parent(s): e0e71cb

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +5 -14
README.md CHANGED
@@ -2,22 +2,16 @@
2
  license: cc-by-4.0
3
  pretty_name: Pixels Point Polygons
4
  size_categories:
5
- - 100K<n<1M
6
  task_categories:
7
  - image-segmentation
8
- - object-detection
9
  tags:
 
10
  - Aerial
 
11
  - Environement
12
  - Multimodal
13
- - Building
14
- - Polygon
15
- - Vectorization
16
- - LIDAR
17
- - ALS
18
- - Image
19
- language:
20
- - en
21
  ---
22
 
23
 
@@ -32,9 +26,6 @@ language:
32
  <b>Figure 1</b>: A view of our dataset of Zurich, Switzerland
33
  </div>
34
 
35
-
36
- <!-- [[Project Webpage]()] [[Paper](https://arxiv.org/abs/2412.07899)] [[Video]()] -->
37
-
38
  ## Abstract:
39
 
40
  We present the P<sup>3</sup> dataset, a large-scale multimodal benchmark for building vectorization, constructed from aerial LiDAR point clouds, high-resolution aerial imagery, and vectorized 2D building outlines, collected across three continents. The dataset contains over 10 billion LiDAR points with decimeter-level accuracy and RGB images at a ground sampling distance of 25 cm. While many existing datasets primarily focus on the image modality, P$^3$ offers a complementary perspective by also incorporating dense 3D information. We demonstrate that LiDAR point clouds serve as a robust modality for predicting building polygons, both in hybrid and end-to-end learning frameworks. Moreover, fusing aerial LiDAR and imagery further improves accuracy and geometric quality of predicted polygons. The P<sup>3</sup> dataset is publicly available, along with code and pretrained weights of three state-of-the-art models for building polygon prediction at https://github.com/raphaelsulzer/PixelsPointsPolygons.
@@ -147,4 +138,4 @@ This repository benefits from the following open-source work. We thank the autho
147
 
148
  1. [Frame Field Learning](https://github.com/Lydorn/Polygonization-by-Frame-Field-Learning)
149
  2. [HiSup](https://github.com/SarahwXU/HiSup)
150
- 3. [Pix2Poly](https://github.com/yeshwanth95/Pix2Poly)
 
2
  license: cc-by-4.0
3
  pretty_name: Pixels Point Polygons
4
  size_categories:
5
+ - 10B<n<100B
6
  task_categories:
7
  - image-segmentation
 
8
  tags:
9
+ - IGN
10
  - Aerial
11
+ - Satellite
12
  - Environement
13
  - Multimodal
14
+ - Earth Observation
 
 
 
 
 
 
 
15
  ---
16
 
17
 
 
26
  <b>Figure 1</b>: A view of our dataset of Zurich, Switzerland
27
  </div>
28
 
 
 
 
29
  ## Abstract:
30
 
31
  We present the P<sup>3</sup> dataset, a large-scale multimodal benchmark for building vectorization, constructed from aerial LiDAR point clouds, high-resolution aerial imagery, and vectorized 2D building outlines, collected across three continents. The dataset contains over 10 billion LiDAR points with decimeter-level accuracy and RGB images at a ground sampling distance of 25 cm. While many existing datasets primarily focus on the image modality, P$^3$ offers a complementary perspective by also incorporating dense 3D information. We demonstrate that LiDAR point clouds serve as a robust modality for predicting building polygons, both in hybrid and end-to-end learning frameworks. Moreover, fusing aerial LiDAR and imagery further improves accuracy and geometric quality of predicted polygons. The P<sup>3</sup> dataset is publicly available, along with code and pretrained weights of three state-of-the-art models for building polygon prediction at https://github.com/raphaelsulzer/PixelsPointsPolygons.
 
138
 
139
  1. [Frame Field Learning](https://github.com/Lydorn/Polygonization-by-Frame-Field-Learning)
140
  2. [HiSup](https://github.com/SarahwXU/HiSup)
141
+ 3. [Pix2Poly](https://github.com/yeshwanth95/Pix2Poly)