Deep Multi-View Stereo Gone Wild
2021 International Conference on 3D Vision (3DV)(2021)
摘要
Deep multi-view stereo (MVS) methods have been developed and extensively compared on simple datasets, where they now outperform classical approaches. In this paper, we ask whether the conclusions reached in controlled scenarios are still valid when working with Internet photo collections. We propose a methodology for evaluation and explore the influence of three aspects of deep MVS methods: networ...
更多查看译文
关键词
Deep learning,Three-dimensional displays,Systematics,Art,Training data,Network architecture,Internet
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要