Globally Optimal Inlier Set Maximization for Atlanta World Understanding
SCIE
SCOPUS
- Title
- Globally Optimal Inlier Set Maximization for Atlanta World Understanding
- Authors
- Kyungdon Joo; Tae-Hyun Oh; In So Kweon; Jean-Charles Bazin
- Date Issued
- 2020-10
- Publisher
- Institute of Electrical and Electronics Engineers
- Abstract
- Status: Accepted
In this work, we describe man-made structures via an appropriate structure assumption, called Atlanta world, which
contains a vertical direction (typically the gravity direction) and a set of horizontal directions orthogonal to the vertical direction.
Contrary to the commonly used Manhattan world assumption, the horizontal directions in Atlanta world are not necessarily orthogonal
to each other. While Atlanta world permits to encompass a wider range of scenes, this makes the search space much larger and the
problem more challenging. Our input data is a set of surface normals, for example acquired from RGB-D cameras or 3D laser scanners,
and also lines from calibrated images. Given this input data, we propose the first globally optimal method of inlier set maximization for
Atlanta direction estimation. We define a novel search space for Atlanta world, as well as its parametrization, and solve this challenging
problem by a branch-and-bound (BnB) framework. To alleviate the computational bottleneck in BnB, i.e. bound computation, we present
two bound computation strategies: rectangular bound and slice bound in an efficient measurement domain, i.e. the extended Gaussian
image (EGI). In addition, we propose an efficient two-stage method to automatically estimate the number of horizontal directions of a
scene. Experimental results with synthetic and real-world datasets have successfully confirmed the validity of our approach.
- URI
- https://oasis.postech.ac.kr/handle/2014.oak/103516
- DOI
- 10.1109/TPAMI.2019.2909863
- ISSN
- 0162-8828
- Article Type
- Article
- Citation
- IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 42, no. 10, page. 2656 - 2669, 2020-10
- Files in This Item:
- There are no files associated with this item.
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.