arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1910.04099cs.CVcs.LGeess.IV
收藏

基于单幅360图像的曼哈顿房间布局重建:几种先进方法的比较研究

Manhattan Room Layout Reconstruction from a Single 360 image: A Comparative Study of State-of-the-art Methods

University of Illinois at Urbana-Champaign(伊利诺伊大学厄巴纳-香槟分校);National Tsing Hua University(国立清华大学);National Chiao Tung University(国立交通大学);ShanghaiTech University(上海科技大学);University of Washington(华盛顿大学);Apple Inc.(苹果公司);King Abdullah University of Science and Technology(阿卜杜拉国王科技大学)

Chuhang Zou, Jheng-Wei Su, Chi-Han Peng, Alex Colburn, Qi Shan, Peter Wonka, Hung-Kuo Chu, Derek Hoiem

摘要:

最近的方法,预测布局从360度全景图产生了很好的结果。这些方法建立在由三个步骤组成的共同框架上:基于基于边缘对齐的预处理步骤、布局元素的预测以及通过将3D布局拟合到布局元素的后处理步骤。到目前为止,由于多个不同的设计决策,例如编码网络(例如SegNet或ResNet),预测的元素类型(例如角落,墙壁/地板边界或语义分割)或拟合3D布局的方法,很难比较这些方法。为了应对这一挑战,我们总结和描述的共同框架,变量,和设计决策的影响。为了进行完整的评估,我们还提出了Matterport 3D数据集的扩展注释[3],并引入了两个基于深度的评估指标。

英文摘要:

Recent approaches for predicting layouts from 360 panoramas produce excellent results. These approaches build on a common framework consisting of three steps: a pre-processing step based on edge-based alignment, prediction of layout elements, and a post-processing step by fitting a 3D layout to the layout elements. Until now, it has been difficult to compare the methods due to multiple different design decisions, such as the encoding network (e.g. SegNet or ResNet), type of elements predicted (e.g. corners, wall/floor boundaries, or semantic segmentation), or method of fitting the 3D layout. To address this challenge, we summarize and describe the common framework, the variants, and the impact of the design decisions. For a complete evaluation, we also propose extended annotations for the Matterport3D dataset [3], and introduce two depth-based evaluation metrics.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑