Abstract
The rotation averaging problem is a fundamental task in computer vision applications. It is generally very difficult to solve due to the nonconvex rotation constraints. While a sufficient optimality condition is available in the literature, there is a lack of a fast convergent algorithm to achieve stationary points. In this paper, by exploring the problem structure, we first propose a block coordinate descent (BCD)-based rotation averaging algorithm with guaranteed convergence to stationary points. Afterwards, we further propose an alternative rotation averaging algorithm by applying successive upper-bound minimization (SUM) method. The SUM-based rotation averaging algorithm can be implemented in parallel and thus is more suitable for addressing large-scale rotation averaging problems. Numerical examples verify that the proposed rotation averaging algorithms have superior convergence performance as compared to the state-of-the-art algorithm. Moreover, by checking the sufficient optimality condition, we find from extensive numerical experiments that the proposed two algorithms can achieve globally optimal solutions.
- Alexandr Andoni, Piotr Indyk, Thijs Laarhoven, Ilya Razenshteyn, and Ludwig Schmidt. 2015. Practical and optimal LSH for angular distance. In Advances in Neural Information Processing Systems. 1225--1233.Google Scholar
- Dimitri P Bertsekas. 1997. Nonlinear programming. Journal of the Operational Research Society 48, 3 (1997), 334--334.Google Scholar
Cross Ref
- Yuchao Dai, Jochen Trumpf, Hongdong Li, Nick Barnes, and Richard Hartley. 2009. Rotation averaging with application to camera-rig calibration. In Asian Conference on Computer Vision. Springer, 335--346.Google Scholar
- Olof Enqvist, Fredrik Kahl, and Carl Olsson. 2011. Non-sequential structure from motion. In 2011 IEEE International Conference on Computer Vision Workshops (ICCV Workshops). IEEE, 264--271.Google Scholar
Cross Ref
- Anders Eriksson, Carl Olsson, Fredrik Kahl, and Tat-Jun Chin. 2018. Rotation averaging and strong duality. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 127--135.Google Scholar
Cross Ref
- Johan Fredriksson and Carl Olsson. 2012. Simultaneous multiple rotation averaging using lagrangian duality. In Asian Conference on Computer Vision. Springer, 245--258.Google Scholar
- Venu Madhav Govindu. 2001. Combining two-view constraints for motion estimation. In Proceedings of the 2001 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. CVPR 2001, Vol. 2. IEEE, II--II.Google Scholar
Cross Ref
- Andrew J Hanson. 2005. Visualizing quaternions. In ACM SIGGRAPH 2005 Courses. ACM, 1.Google Scholar
Digital Library
- Richard Hartley, Jochen Trumpf, Yuchao Dai, et al. 2010. Rotation averaging and weak convexity. In Proc. of the 19th International Symposium on Mathematical Theory of Networks and Systems (MTNS'10). 2435--2442.Google Scholar
- Richard Hartley, Jochen Trumpf, Yuchao Dai, and Hongdong Li. 2013. Rotation averaging. International journal of computer vision 103, 3 (2013), 267--305.Google Scholar
Cross Ref
- Jan J Koenderink and Andrea J Van Doorn. 1991. Affine structure from motion. JOSA A 8, 2 (1991), 377--385.Google Scholar
Cross Ref
- Daniel Martinec and Tomas Pajdla. 2007. Robust rotation and translation estimation in multiview reconstruction. In 2007 IEEE Conference on Computer Vision and Pattern Recognition. IEEE, 1--8.Google Scholar
Cross Ref
- Pauline C Ng and Steven Henikoff. 2003. SIFT: Predicting amino acid changes that affect protein function. Nucleic acids research 31, 13 (2003), 3812--3814.Google Scholar
- Sverker Rasmuson, Erik Sintorn, and Ulf Assarsson. 2020. User-guided 3D reconstruction using multi-view stereo. In Symposium on Interactive 3D Graphics and Games. 1--9.Google Scholar
Digital Library
- Meisam Razaviyayn, Mingyi Hong, and Zhi-Quan Luo. 2013. A unified convergence analysis of block successive minimization methods for nonsmooth optimization. SIAM Journal on Optimization 23, 2 (2013), 1126--1153.Google Scholar
Cross Ref
- Denise Sakai. 1994. Labeling chordal graphs: distance two condition. SIAM Journal on Discrete Mathematics 7, 1 (1994), 133--140.Google Scholar
Digital Library
- Noah Snavely, Steven M Seitz, and Richard Szeliski. 2006. Photo tourism: exploring photo collections in 3D. In ACM transactions on graphics (TOG), Vol. 25. ACM, 835--846.Google Scholar
- Camillo J Taylor and David J Kriegman. 1994. Minimization on the Lie group SO (3) and related manifolds. Yale University 16, 155 (1994), 6.Google Scholar
- Kyle Wilson, David Bindel, and Noah Snavely. 2016. When is rotations averaging hard?. In European Conference on Computer Vision. Springer, 255--270.Google Scholar
Cross Ref
- Stephen J Wright. 2015. Coordinate descent algorithms. Mathematical Programming 151, 1 (2015), 3--34.Google Scholar
Digital Library
- Yihong WU Zhanyi Hu, Fuchao Wu and Qiulei Dong. [n.d.]. 3D reconstruction dataset. http://vision.ia.ac.cn/data.Google Scholar
Index Terms
Efficient Algorithms for Rotation Averaging Problems
Recommendations
Globally Optimal Estimates for Rotation Averaging Problems
IHMSC '14: Proceedings of the 2014 Sixth International Conference on Intelligent Human-Machine Systems and Cybernetics - Volume 02This paper investigates the difficulty of obtaining the global optimum in rotation averaging problems and presents different representations of rotation matrix. One of the drawbacks is that there is a 2-to-1 mapping when using the angle-axis and ...
Incremental Rotation Averaging
AbstractIn this paper, we present a simple yet effective rotation averaging pipeline, termed Incremental Rotation Averaging (IRA), which is inspired by the well-developed incremental Structure from Motion (SfM) techniques. Unlike the traditional rotation ...
A First-Order Primal-Dual Algorithm for Convex Problems with Applications to Imaging
In this paper we study a first-order primal-dual algorithm for non-smooth convex optimization problems with known saddle-point structure. We prove convergence to a saddle-point with rate O (1/ N ) in finite dimensions for the complete class of ...






Comments