期刊
2019 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2019)
卷 -, 期 -, 页码 5951-5960出版社
IEEE COMPUTER SOC
DOI: 10.1109/CVPR.2019.00611
关键词
-
资金
- SNF [200021_162958]
- ISF [2366/16]
- NSFC [61761146002]
- LHTD [20170003]
- National Engineering Laboratory for Big Data System Computing Technology
- Swiss National Science Foundation (SNF) [200021_162958] Funding Source: Swiss National Science Foundation (SNF)
We present a detail-driven deep neural network for point set upsampling. A high-resolution point set is essential for point-based rendering and surface reconstruction. Inspired by the recent success of neural image super-resolution techniques, we progressively train a cascade of patch-based upsampling networks on different levels of detail end-to-end. We propose a series of architectural design contributions that lead to a substantial performance boost. The effect of each technical contribution is demonstrated in an ablation study. Qualitative and quantitative experiments show that our method significantly outperforms the state-of-the- art learning-based [58, 59], and optimazation-based [23] approaches, both in terms of handling low-resolution inputs and revealing high-fidelity details. The data and code are at http://githup.com/yifita/3pu.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据