Abstract
Robustly fitting a linear model from outlier-contaminated data is an important and basic task in many scientific fields, and it is often tackled by consensus set maximization. There have been several studies on globally optimal methods for consensus set maximization, but most of them are currently confined to problems with small number of input observations and low outlier ratios. In this paper, we develop a globally optimal algorithm aiming at consensus set maximization to solve the robust linear model fitting problems with the unit-norm constraint, which is based on the branch-and-bound optimization framework. The unit-norm constraint is utilized to fix the unknown scale of linear model parameters, and we propose a compact representation of the unit-bounded searching domain to avoid introducing the additional non-linearity in the unit-norm constraint. The compact representation leads to a geometrically derived bound, which accelerates the calculation and enables the method to handle the problems with large number of observations. Experiments on both synthetic and real data show that the proposed algorithm outperforms existing globally optimal methods, especially in low dimensional problems with large number of input observations and high outlier ratios. The implementation of the source code is publicly available https://github.com/YiruWangYuri/Demo-for-GoCR.
Similar content being viewed by others
References
Adjiman, C. S., Androulakis, I. P., & Floudas, C. A. (1998). A global optimization method, \(\alpha \)bb, for general twice-differentiable constrained NLPs-II. implementation and computational results. Computers & Chemical Engineering, 22(9), 1159–1179.
Aftab, K., Hartley, R., & Trumpf, J. (2014). Generalized Weiszfeld algorithms for LQ optimization. IEEE Transactions on Pattern Analysis and Machine Intelligence, 37(4), 728–745.
Barron, J.T. (2019). A general and adaptive robust loss function. In: Computer vision and pattern recognition (pp. 4331–4339). IEEE.
Bazin, J.C., Seo, Y., Demonceaux, C., Vasseur, P., Ikeuchi, K., Kweon, I., & Pollefeys, M. (2012). Globally optimal line clustering and vanishing point estimation in Manhattan world. In: Computer vision and pattern recognition (pp. 638–645). IEEE.
Bazin, J. C., Li, H., Kweon, I. S., Demonceaux, C., Vasseur, P., & Ikeuchi, K. (2012). A branch-and-bound approach to correspondence and grouping problems. IEEE Transactions on Pattern Analysis and Machine Intelligence, 35(7), 1565–1576.
Benedek, C., Sziranyi, T., Kato, Z., & Zerubia, J. (2009). Detection of object motion regions in aerial image pairs with a multilayer markovian model. IEEE Transactions on Image Processing, 18(10), 2303–2315.
Brachmann, E., & Rother, C. (2019). Neural-guided RANSAC: Learning where to sample model hypotheses. In: International conference on computer vision (pp. 4322–4331). IEEE.
Brachmann, E., Krull, A., Nowozin, S., Shotton, J., Michel, F., Gumhold, S., & Rother, C. (2017). DSAC-differentiable RANSAC for camera localization. In: Computer vision and pattern recognition (pp. 6684–6692). IEEE.
Cai, Z., Chin, T.J., & Koltun, V.(2019). Consensus maximization tree search revisited. In: International conference on computer vision (pp. 1637–1645). IEEE.
Cai, Z., Chin, T.J., Le, H., Suter, D.(2018).Deterministic consensus maximization with biconvex programming. In: European conference on computer vision (pp. 685–700). Springer.
Chin, T.J., Cai, Z., & Neumann, F.(2018). Robust fitting in computer vision: Easy or hard? In: European conference on computer vision (pp. 701–716). Springer.
Chin, T.J., HengKee, Y., Eriksson, A., & Neumann, F. (2016). Guaranteed outlier removal with mixed integer linear programs. In: Computer vision and pattern recognition (pp. 5858–5866). IEEE.
Chin, T.J., Suter, D., Medioni, G., & Dickinson, S. (2017). The maximum consensus problem: Recent algorithmic advances. Morgan & Claypool Publishers.
Chin, T. J., Purkait, P., Eriksson, A., & Suter, D. (2016). Efficient globally optimal consensus maximisation with tree search. IEEE Transactions on Pattern Analysis and Machine Intelligence, 39(4), 758–772.
Choi, S., Kim, T., & Yu, W. (2009). Performance evaluation of Ransac family. In: British machine vision conference (pp. 1–12).
Enqvist, O., Ask, E., Kahl, F., & Åström, K. (2012). Robust fitting for multiple view geometry. In: European conference on computer vision (pp. 738–751). Springer.
Enqvist, O., Ask, E., Kahl, F., & Åström, K. (2015). Tractable algorithms for robust model estimation. International Journal of Computer Vision, 112(1), 115–129.
Fischler, M. A., & Bolles, R. C. (1981). Random sample consensus: A paradigm for model fitting with applications to image analysis and automated cartography. Communications of the ACM, 24(6), 381–395.
Fredriksson, J., Larsson, V., Olsson, C., Enqvist, O., & Kahl, F. (2016). Efficient algorithms for robust estimation of relative translation. Image and Vision Computing, 52, 114–124.
Hartley, R. I., & Kahl, F. (2009). Global optimization through rotation space search. International Journal of Computer Vision, 82(1), 64–79.
Hartley, R., Trumpf, J., Dai, Y., & Li, H. (2013). Rotation averaging. International Journal of Computer Vision, 103(3), 267–305.
Hartley, R., & Zisserman, A. (2003). Multiple view geometry in computer vision. Cambridge University Press.
Heller, J., Havlena, M., & Pajdla, T. (2016). Globally optimal hand-eye calibration using branch-and-bound. IEEE Transactions on Pattern Analysis and Machine Intelligence, 38(5), 1027–1033.
Holland, P. W., & Welsch, R. E. (1977). Robust regression using iteratively reweighted least-squares. Communications in Statistics-theory and Methods, 6(9), 813–827.
http://mplab.sztaki.hu/remotesensing/airmotion_benchmark.html
Ikami, D., Yamasaki, T., & Aizawa, K. (2018). Fast and robust estimation for unit-norm constrained linear fitting problems. In: Computer vision and pattern recognition (pp. 8147–8155). IEEE.
Inkilä, K. (2005). Homogeneous least squares problem. Photogrammetric Journal of Finland, 19(2), 34–42.
John, C. D., Lars, P., Laurent, K., & Li, H. (2020). Globally-optimal inlier set maximisation for camera pose and correspondence estimation. IEEE Transactions on Pattern Analysis and Machine Intelligence, 42(2), 328–342.
Joo, K., Oh, T. H., Kim, J., & Kweon, I. S. (2018). Robust and globally optimal manhattan frame estimation in near real time. IEEE Transactions on Pattern Analysis and Machine Intelligence, 41(3), 682–696.
Le, H., Chin, T.J., & Suter, D. (2017). Ratsac-random tree sampling for maximum consensus estimation. In: Digital image computing: techniques and applications (pp. 1–8). IEEE.
Le, H.M., Do, T.T., Hoang, T., & Cheung, N.M. (2019). SDRSAC: Semidefinite-based randomized approach for robust point cloud registration without correspondences. In: Computer vision and pattern recognition (pp. 124–133). IEEE.
Le, H. M., Chin, T. J., Eriksson, A., Do, T. T., & Suter, D. (2021). Deterministic approximate methods for maximum consensus robust fitting. IEEE Transactions on Pattern Analysis and Machine Intelligence, 43(3), 842–857.
Li, H. (2009). Consensus set maximization with guaranteed global optimality for robust geometry estimation. In: International conference on computer vision (pp. 1074–1080). IEEE.
Liu, Y., Dong, Y., Song, Z., & Wang, M. (2018). 2D–3D Point Set Registration Based on Global Rotation Search. IEEE Transactions on Image Processing, 28(5), 2599–2613.
Moore, R.E., Kearfott, R.B., & Cloud, M.J. (2009). Introduction to interval analysis. SIAM.
Morrison, D. R., Jacobson, S. H., Sauppe, J. J., & Sewell, E. C. (2016). Branch-and-bound algorithms: A survey of recent advances in searching, branching, and pruning. Discrete Optimization, 19, 79–102.
Oja, E., & Wang, L. (1996). Robust fitting by nonlinear neural units. Neural Networks, 9(3), 435–444.
Olsson, C., Enqvist, O., & Kahl, F. (2008). A polynomial-time bound for matching and registration with outliers. In: Computer vision and pattern recognition (pp. 1–8). IEEE.
Raguram, R., Chum, O., Pollefeys, M., Matas, J., & Frahm, J. M. (2012). USAC: A universal framework for random sample consensus. IEEE Transactions on Pattern Analysis and Machine Intelligence, 35(8), 2022–2038.
Ruckstuhl, A. (2014). Robust fitting of parametric models based on m-estimation. Lecture notes.
Speciale, P., PaniPaudel, D., Oswald, M.R., Kroeger, T., Van Gool, L., & Pollefeys, M. (2017).Consensus maximization with linear matrix inequality constraints. In: Computer vision and pattern recognition (pp. 4941–4949). IEEE.
Tennakoon, R. B., Bab-Hadiashar, A., Cao, Z., Hoseinnezhad, R., & Suter, D. (2015). Robust model fitting using higher than minimal subset sampling. IEEE Transactions on Pattern Analysis and Machine Intelligence, 38(2), 350–362.
Tzoumas, V., Antonante, P., & Carlone, L. (2019).Outlier-robust spatial perception: Hardness, general-purpose algorithms, and guarantees. In: Intelligent robots and systems (pp. 5383–5390). IEEE.
Wang, Y., Liu, Y., Li, X., Wang, C., Wang, M., & Song, Z. (2021). Practical globally optimal consensus maximization by branch-and-bound based on interval arithmetic. Pattern Recognition, 115, 107897.
Yang, H., & Carlone, L. (2019). A quaternion-based certifiably optimal solution to the Wahba problem with outliers. In: International conference on computer vision. Springer.
Yang, J., Li, H., Campbell, D., & Jia, Y. (2015). Go-ICP: A globally optimal solution to 3D ICP point-set registration. IEEE Transactions on Pattern Analysis and Machine Intelligence, 38(11), 2241–2254.
Zach, C., & Bourmaud, G. (2017). Iterated lifting for robust cost optimization. In: British machine vision conference.
Zach, C., & Bourmaud, G. (2018). Descending, lifting or smoothing: Secrets of robust cost optimization. In: European conference on computer vision (pp. 547–562). Springer.
Zheng, Y., Sugimoto, S., & Okutomi, M.(2011).Deterministically maximizing feasible subsystem for robust model fitting with unit norm constraint. In: Computer vision and pattern recognition (pp. 1825–1832). IEEE.
Zhou, Q.Y., Park, J., & Koltun, V. (2016). Fast global registration. In: European conference on computer vision (pp. 766–782). Springer.
Acknowledgements
This work was supported by National Natural Science Foundation of China under Grant 82072021, by Shanghai Municipal Science and Technology Major Project (No. 2018SHZDZX01), by Shanghai Rising Star Program (No. 21QC1400900), by the European Union’s Horizon 2020 Framework Programme for Research and Innovation under the Specic Grant Agreement No. 945539 (Human Brain Project SGA3) and by National Natural Science Foundation of China under Grant 62076070.
Author information
Authors and Affiliations
Corresponding authors
Ethics declarations
Conflicts of interest
The authors declare that they have no conflict of interest.
Additional information
Communicated by Akihiro Sugimoto.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Information
Below is the link to the electronic supplementary material.
Supplementary file 1 (mp4 3277 KB)
Appendix: Proof of Lemma 1
Appendix: Proof of Lemma 1
Lemma 2
\(\varvec{x}_p\in \mathbb {S}^{d+} \), \(\varvec{x}_q \in \mathbb {S}^{d+}\) are two vectors in the d-dimensional unit hyper-hemisphere, and \(\varvec{\varphi }_p \in \mathbb {R}^d\), \(\varvec{\varphi }_q \in \mathbb {R}^d\) are the two corresponding points in the d-dimensional solid ball.
Proof
\(\varvec{x}_p, \varvec{x}_q, \varvec{\varphi }_p, \varvec{\varphi }_q\) can be expressed as:
Since \(\Vert \varvec{x}_p \Vert = \Vert \varvec{x}_q \Vert = \Vert \varvec{\omega }_p \Vert = \Vert \varvec{\omega }_q \Vert = 1\), we can define \(\gamma =\angle (\varvec{\omega }_p,\varvec{\omega }_q)=\arccos (\frac{\varvec{\omega _p^T\omega _q}}{\Vert \varvec{\omega }_p \Vert \Vert \varvec{\omega }_q \Vert } )=\arccos (\varvec{\omega _p^T\omega _q})\).
On the left, we have
Let \(\lambda \in [0,1]\), \(\cos (\gamma )= 2\lambda -1\), then we have
On the right,
Since \(\cos (\gamma )= 2\lambda -1\),
To compare \(\rho \) and \(\xi \), we introduce a convex function \(f(\alpha )=\arccos ^2(\alpha )\), when \(\alpha \in [-1,1]\), see Fig. 7. According to Jensen’s inequality, we then have
Let \(\alpha _1=\cos (\theta _p-\theta _q)\) and \(\alpha _2=\cos (\theta _p+\theta _q)\), then we have
Because of the non-negativeness of \(\rho \) and \(\xi \), then \(\angle (\varvec{x}_p,\varvec{x}_q)\le \Vert \varvec{\varphi }_p-\varvec{\varphi }_q\Vert \) is finally proved.
Rights and permissions
About this article
Cite this article
Liu, Y., Wang, Y., Wang, M. et al. Globally Optimal Linear Model Fitting with Unit-Norm Constraint. Int J Comput Vis 130, 933–946 (2022). https://doi.org/10.1007/s11263-022-01574-z
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11263-022-01574-z