Update README.md
Browse files
README.md
CHANGED
|
@@ -2,4 +2,17 @@
|
|
| 2 |
license: mit
|
| 3 |
datasets:
|
| 4 |
- ShapeNet/ShapeNetCore
|
| 5 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2 |
license: mit
|
| 3 |
datasets:
|
| 4 |
- ShapeNet/ShapeNetCore
|
| 5 |
+
title: Transformer-Based Point Cloud Denoising/Completion Models
|
| 6 |
+
---
|
| 7 |
+
|
| 8 |
+
Four variants of the model architecture were trained on a random subset (50%) of ShapeNetCore.
|
| 9 |
+
The models are based on two types of data augmentations: random noise perturbations (Gaussian normal) and point deletion (using uniform sampling). The models are as follow:
|
| 10 |
+
|
| 11 |
+
|
| 12 |
+
| Model | noise strength (%) | points removed (%) |
|
| 13 |
+
| :---------------- | :------: | :----: |
|
| 14 |
+
| pc_denoiser_1 | 1 | 0 |
|
| 15 |
+
| pc_denoiser_2 | 1 | 50 |
|
| 16 |
+
| pc_denoiser_3 | 2 | 0 |
|
| 17 |
+
| pc_denoiser_4 | 2 | 50 |
|
| 18 |
+
|