PointNorm-Net: Self-Supervised Normal Prediction of 3D Point Clouds via Multi-Modal Distribution Estimation
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2023
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915234077212672 |
|---|---|
| author | Zhang, Jie Nie, Minghui Zou, Changqing Liu, Jian Liu, Ligang Cao, Junjie |
| author_facet | Zhang, Jie Nie, Minghui Zou, Changqing Liu, Jian Liu, Ligang Cao, Junjie |
| contents | Although supervised deep normal estimators have recently shown impressive results on synthetic benchmarks, their performance deteriorates significantly in real-world scenarios due to the domain gap between synthetic and real data. Building high-quality real training data to boost those supervised methods is not trivial because point-wise annotation of normals for varying-scale real-world 3D scenes is a tedious and expensive task. This paper introduces PointNorm-Net, the first self-supervised deep learning framework to tackle this challenge. The key novelty of PointNorm-Net is a three-stage multi-modal normal distribution estimation paradigm that can be integrated into either deep or traditional optimization-based normal estimation frameworks. Extensive experiments show that our method achieves superior generalization and outperforms state-of-the-art conventional and deep learning approaches across three real-world datasets that exhibit distinct characteristics compared to the synthetic training data. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2304_04884 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | PointNorm-Net: Self-Supervised Normal Prediction of 3D Point Clouds via Multi-Modal Distribution Estimation Zhang, Jie Nie, Minghui Zou, Changqing Liu, Jian Liu, Ligang Cao, Junjie Computer Vision and Pattern Recognition Although supervised deep normal estimators have recently shown impressive results on synthetic benchmarks, their performance deteriorates significantly in real-world scenarios due to the domain gap between synthetic and real data. Building high-quality real training data to boost those supervised methods is not trivial because point-wise annotation of normals for varying-scale real-world 3D scenes is a tedious and expensive task. This paper introduces PointNorm-Net, the first self-supervised deep learning framework to tackle this challenge. The key novelty of PointNorm-Net is a three-stage multi-modal normal distribution estimation paradigm that can be integrated into either deep or traditional optimization-based normal estimation frameworks. Extensive experiments show that our method achieves superior generalization and outperforms state-of-the-art conventional and deep learning approaches across three real-world datasets that exhibit distinct characteristics compared to the synthetic training data. |
| title | PointNorm-Net: Self-Supervised Normal Prediction of 3D Point Clouds via Multi-Modal Distribution Estimation |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2304.04884 |