Abstract
Computing the Riemannian center of mass or the finite sample Fréchet mean has attracted enormous attention lately due to the easy availability of data that are manifold valued. Manifold-valued data are encountered in numerous domains including but not limited to medical image computing, mechanics, statistics, machine learning. It is common practice to estimate the finite sample Fréchet mean by using a gradient descent technique to find the minimum of the Fréchet function when it exists. The convergence rate of this gradient descent method depends on many factors including the step size and the variance of the given manifold-valued data. As an alternative to the gradient descent technique, we propose a recursive (incremental) algorithm for estimating the Fréchet mean/expectation (iFEE) of the distribution from which the sample data are drawn. The proposed algorithm can be regarded as a geometric generalization of the well-known incremental algorithm for computing arithmetic mean, since it reinterprets this algebraic formula in terms of geometric operations on geodesics in the more general manifold setting. In particular, given known formulas for geodesics, iFEE does not require any optimization in contrast to the non-incremental counterparts and offers significant improvement in efficiency and flexibility. For the case of simply connected, complete and nonpositively curved Riemannian manifolds, we prove that iFEE converges to the true expectation in the limit. We present several experiments demonstrating the efficiency and accuracy of iFEE in comparison to the non-incremental counterpart for computing the finite sample Fréchet mean of symmetric positive definite matrices as well as applications of iFEE to K-means clustering and diffusion tensor image segmentation.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
This topological assumption is not particularly restrictive since by Cartan–Hadamard Theorem [7], any simply connected complete d-dimensional Riemannian manifold with nonpositive sectional curvature can be constructed topologically from \(\mathbb{R}^{d}\).
- 2.
Points in \(\mathcal{M}\) are denoted by boldface font and their corresponding points in the tangent space are denoted by regular font.
References
Afsari B, Tron R, Vidal R (2013) On the convergence of gradient descent for finding the riemannian center of mass. SIAM J Control Optim 51(3):2230–2260
Amari S (2001) Information geometry. American Mathematical Society, Providence
Ando T, Li CK, Mathias R (2004) Geometric means. Linear Algebra Appl 385(2):305–334
Arsigny V, Fillard P, Pennec X, Ayache N (2006) Log-Euclidean metrics for fast and simple calculus on diffusion tensors. Magn Reson Med 56:411–421
Barmpoutis A, Vemuri BC, Shepherd TM, Forder JR (2007) Tensor splines for interpolation and approximation of DT-MRI with applications to segmentation of isolated rat hippocampi. IEEE Trans Med Imaging 26:1537–1546
Basser PJ, Mattiello J, Lebihan D (1994) Estimation of the effective self-diffusion tensor from NMR spin echo. J Magn Reson 103(3):247–254
Berger M (2007) A panoramic view of Riemannian geometry. Springer, Berlin
Bhatia R, Holbrook J (2006) Riemannian geometry and matrix geometric means. Linear Algebra Appl 413(2):594–618
Bhattacharya R, Patrangenaru V (2003) Large sample theory of intrinsic and extrinsic sample means on manifolds - I. Ann Stat 31(1):1–29
Bhattacharya R, Patrangenaru V (2005) Large sample theory of intrinsic and extrinsic sample means on manifolds - II. Ann Stat 33(3):1225–1259
Bini DA, Meini B, Poloni F (2010) An effective matrix geometric mean satisfying the ando-li-mathias properties. Math Comput 79(269):437–452
Bridson M, Haefliger (1999) A metric spaces of non-positive curvature. Springer, Berlin
Caselles V, Kimmel R, Sapiro G (1997) Geodesic active contours. Int J Comput Vis 22 (1):61–79
Cetingul H, Vidal R (2009) Intrinsic mean shift for clustering on Stiefel and Grassmann manifolds. In: Proceedings of IEEE international conference on computer vision and pattern recognition, pp 1896–1902
Chan T, Vese L (2001) Active contours without edges. IEEE Trans Image Process 10(2): 266–277
Cheeger J, Ebin D (2008) Comparison theorems in Riemannian geometry. American Mathematical Society, Providence
Cheng G, Vemuri BC (2013) A novel dynamic system in the space of SPD matrices with applications to appearance tracking. SIAM J Imag Sci 6(1):592–615
Cheng G, Salehian H, Vemuri BC (2012) Efficient recursive algorithms for computing the mean diffusion tensor and applications to DTI segmentation. In: ECCV, vol 7, pp 390–401
Doretto G, Chiuso A, Wu YN, Soatto S (2003) Dynamic textures. Int J Comput Vis 51(2): 91–109
Feddern C, Weickert J, Burgeth B (2003) Level-set methods for tensor-valued images. In: Proceedings of 2nd IEEE workshop on variational, geometric and level set methods in computer vision, pp 65–72
Fletcher T, Joshi S (2004) Principal geodesic analysis on symmetric spaces: statistics of diffusion tensors. In: Computer vision and mathematical methods in medical and biomedical image analysis, pp 87–98
Fletcher P, Joshi S (2007) Riemannian geometry for the statistical analysis of diffusion tensor data. Signal Process 87(2):250–262
Fréchet M (1948) Les éléments aléatoires de nature quelconque dans un espace distancié. Ann Inst Henri Poincare 10(4):215–310
Ginestet C (2012) Strong and weak laws of large numbers for Frechet sample means in bounded metric spaces. arXiv:1204.3183
Hartley R, Trumpf J, Dai Y, Li H (2013) Rotation averaging. Int J Comput Vis 103(3):267–305
Helgason S (2001) Differential geometry, lie groups, and symmetric spaces. American Mathematical Society, Providence
Ho J, Xie Y, Vemuri BC (2013) On A nonlinear generalization of sparse coding and dictionary learning. In: ICML, pp 1480–1488
Ho J, Cheng G, Salehian H, Vemuri B (2013) Recursive karcher expectation estimators and geometric law of large numbers. In: Proceedings of the 16th international conference on artificial intelligence and statistics, pp 325–332
Horn B (1886) Robot vision. MIT Press, Cambridge
Kendall D (1984) Shape manifolds, procrustean metrics, and complex projective spaces. Bull Lond Math Soc 16:18–121
Kendall W, Le H (2011) Limit theorems for empirical Frechet means of independent and non-identically distributed manifold-valued random variables. Braz J Probab Stat 25(3):323–352
Khoshnevisan D (2007) Probability. Graduate studies in mathematics, vol 80. American Mathematical Society, Providence
Le H (2001) Locating Féchet means with application to shape spaces. Adv Appl Probab 33(2):324–338
Lenglet C, Rousson M, Deriche R, Faugeras O (2006) Statistics on the manifold of multivariate normal distributions: theory and application to diffusion tensor MRI processing. J Math Imaging Vision 25:423–444. doi:10.1007/s10851-006-6897-z. http://www.dx.doi.org/10.1007/s10851-006-6897-z
Levy A, Lindenbaum M (2000) Sequential karhunen–loeve basis extraction and its application to images. IEEE Trans Image Process 9(8):1371–1374
Lim Y, Pálfia M (2014) Weighted inductive means. Linear Algebra Appl 453:59–83
Malladi R, Sethian J, Vemuri BC (1995) Shape modeling with front propagation: a level set approach. IEEE Trans Pattern Anal Mach Intell 17(2):158–175
Moakher M (2005) A differential geometric approach to the geometric mean of symmetric positive-definite matrices. SIAM J Matrix Anal Appl 26(3):735–747
Moakher M, Batchelor PG (2006) Symmetric positive-definite matrices: from geometry to applications and visualization. Visualization and processing of tensor fields. Springer, Berlin
Pennec X (2006) Intrinsic statistics on riemannian manifolds: Basic tools for geometric measurements. J. Math. Imaging Vision 25(1):127–154
Pennec X, Fillard P, Ayache N (2006) A Riemannian framework for tensor computing. Int J Comput Vis 66(1):41–66
Ross D, Lim J, Lin RS, Yang MH (2008) A Riemannian framework for tensor computing. Int J Comput Vis 77(1–1):125–141
Schwartzman A (2006) Random ellipsoids and false discovery rates: Statistics for diffusion tensor imaging data. Ph.D. thesis, Stanford University
Sturm KT (2003) Probability measures on metric spaces of nonpositive curvature. In: Auscher P, Coulhon T, Grigoryan A (eds) Heat kernels and analysis on manifolds, graphs, and metric spaces, vol 338. American Mathematical Society, Providence
Subbarao R, Meer P (2009) Nonlinear mean shift over Riemannian manifolds. Int J Comput Vis 84(1):1–20
Sverdrup-Thygeson H (1981) Strong law of large numbers for measures of central tendency and dispersion of random variables in compact metric spaces. Ann Stat 9(1):141–145
Tsai A, Yezzi AJ, Willsky A (2001) Curve evolution implementation of the Mumford-Shah functional for image segmentation, denoising, interpolation, and magnification. IEEE Trans Image Process 10(8):1169–1186
Turaga P, Veeraraghavan A, Srivastava A, Chellappa R (2011) Statistical computations on Grassmann and Stiefel manifolds for image and video-based recognition. IEEE Trans Pattern Anal Mach Intell 33(11):2273–2286
Tuzel O, Porikli F, Meer P (2007) Human detection via classification on Riemannian manifolds. In: Proceedings of IEEE international conference on computer vision and pattern recognition
Tyagi A, Davis J (2008) A recursive filter for linear systems on Riemannian manifolds. In: Proceedings of IEEE international conference on computer vision and pattern recognition
Wang Z, Vemuri B (2004) Tensor field segmentation using region based active contour model. In: European conferene on computer vision (ECCV), pp 304–315
Wang Z, Vemuri B (2005) DTI segmentation using an information theoretic tensor dissimilarity measure. IEEE Trans Med Imaging 24(10):1267–1277
Weldeselassie Y, Hamarneh G (2007) DT-MRI segmentation using graph cuts. In: SPIE Medical Imaging, vol 6512
Wu Y, Wang J, Lu H (2009) Real-time visual tracking via incremental covariance model update on log-euclidean Riemannian manifold. In: CCPR
Xie Y, Vemuri BC, Ho J (2010) Statistical analysis of tensor fields. In: MICCAI, pp 682–689
Ziezold H (1977) On expected figures and a strong law of large numbers for random elements in quasi-metric spaces. In: Transactions of the 7th prague conference on information theory, statistical decision functions, random processes and of the 1974 European meeting of statisticians
Ziyan U, Tuch D, Westin C (2006) Segmentation of thalamic nuclei from DTI using spectral clustering. In: MICCAI, pp 807–814
Acknowledgements
This research was supported in part by the NIH grant NS066340 to BCV.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Appendix
Appendix
In this appendix, we prove Proposition 2.1 presented in Sect. 2.2. The proof is entirely elementary if we assume the general property of CAT(0)-metric spaces [12] and Toponogov’s comparison theorem (specifically, the easier half of the theorem on manifolds of nonpositive sectional curvature) [7].
Complete Riemannian manifolds of nonpositive sectional curvature form an important subclass of CAT(0)-metric spaces [12]. For our purpose, the detailed definition of CAT(0)-metric spaces is not necessary; instead, we will recall only the features that are used in the proof. A geodesic triangle Γ on a complete Riemannian manifold \(\mathcal{M}\) is the union of three geodesic segments joining three points \(\mathbf{p},\mathbf{q},\mathbf{r} \in \mathcal{M}\): \(\gamma _{1}(0) =\gamma _{3}(1) = \mathbf{p},\gamma _{1}(1) =\gamma _{2}(0) = \mathbf{q},\gamma _{2}(1) =\gamma _{3}(0) = \mathbf{r}\). Its comparison triangle Δ is a triangle in \(\mathbb{R}^{2}\) with vertices p, q, r such that the lengths of the sides \(\overline{pq},\overline{qr},\overline{rp}\) are equal to the lengths of the geodesic segments \(\gamma _{1},\gamma _{2},\gamma _{3}\), respectively. Such comparison triangle always exists for any geodesic triangle in \(\mathcal{M}\), and it is unique up to a rigid transform in \(\mathbb{R}^{2}\). The correspondence between the three sides and segments extends naturally to points on the triangles as well: a point x ∈ Γ corresponds to a point x ∈ Δ if their associated sides correspond and their distances to the corresponding two endpoints are the same. For example, if x ∈ γ 1 and \(x \in \overline{pq}\), x corresponds to x if \(\mathbf{d}_{\mathcal{M}}(\mathbf{x},\mathbf{p}) = \mathbf{d}_{\mathbb{R}^{2}}(x,p),\) and hence \(\mathbf{d}_{\mathcal{M}}(\mathbf{x},\mathbf{q}) = \mathbf{d}_{\mathbb{R}^{2}}(x,q)\) as well. An important property enjoyed by any CAT(0)-metric space is that for any pair of points x, y on Γ and their corresponding points x, y on Δ, we have (see Fig. 2.5) \(\mathbf{d}_{\mathcal{M}}(\mathbf{x},\mathbf{y}) \leq \mathbf{d}_{\mathbb{R}^{2}}(x,y).\) The importance of this inequality is the upper bound given by the Euclidean distance in \(\mathbb{R}^{2}\), and it allows us to bound the integral of the squared distance function on \(\mathcal{M}\) by an integral involving squared Euclidean distance that is considerably easier to manage. Finally, for the pair of triangles Γ, Δ, Toponogov’s comparison theorem asserts the angle \(\angle (rpq)\) on Δ no smaller than \(\angle (\mathbf{r}\mathbf{p}\mathbf{q})\) on Γ.
Armed with these results, the proof of Proposition 2.1 is straightforward and it involves comparing two triangles in the tangent space T m . See Fig. 2.1. We restate Proposition 2.1 for convenience below and now present its proof.
Proposition 2.1.
Let x , y , z be two points on a complete Riemannian manifold \(\mathcal{M}\) with nonpositive sectional curvature and γ(t) the unique geodesic path joining x , y such that \(\gamma (0) = \mathbf{x},\gamma (1) = \mathbf{y}\) . Furthermore, let \(x = \mathbf{Log}_{\mathbf{z}}(\mathbf{x}),y = \mathbf{Log}_{\mathbf{z}}(\mathbf{y})\) , and \(\overline{\gamma }(t)\) denote the straight line joining x,y such that \(\overline{\gamma (0)} = x,\overline{\gamma (1)} = y\) . Then, \(\mathbf{d}_{\mathcal{M}}(\gamma (t),\,\mathbf{z}) \leq \|\overline{\gamma (t)}\|.\)
Proof.
Given m k , x k+1 in \(\mathcal{M}\), and m k+1 determined according to the iFEE algorithm, we will denote \(m_{k},x_{k+1}\) and m k+1, their corresponding points in T m under the Riemannian logarithm map Log m . Without loss of generality, we will prove the proposition using \(\mathbf{z} = \mathbf{m},\mathbf{x} = \mathbf{x}_{k+1}\), y = m k . Let \(a = \mathbf{d}_{\mathcal{M}}(\mathbf{x}_{k+1},\mathbf{m})\) and \(b = \mathbf{d}_{\mathcal{M}}(\mathbf{m}_{k},\mathbf{m}).\) On T m , we have the first triangle Σ formed by the three vertices: \(x_{k+1},m_{k}\), and o the origin with the side lengths \(\vert \overline{x_{k+1}\mathbf{o}}\vert = a,\vert \overline{m_{k}\mathbf{o}}\vert = b\). The geodesic triangle Γ on \(\mathcal{M}\) spanned by \(\mathbf{m},\mathbf{x}_{k+1},\mathbf{m}_{k}\) has a comparison triangle Δ in T m spanned by o, p, q with \(\vert \overline{p\mathbf{o}}\vert = a,\vert \overline{q\mathbf{o}}\vert = b\), and by Toponogov’s comparison theorem,
since, by definition, \(\angle (x_{k+1}\mathbf{o}m_{k}) = \angle (\mathbf{x}_{k+1}\mathbf{m}\mathbf{m}_{k})\).
For completing the proof, we need to show that the distance between any point on the side \(\overline{pq}\) of Δ and the origin is not greater than the distance between its corresponding point on the side \(\overline{x_{k+1}m_{k}}\) of Σ and the origin. Specifically, a point \(u \in \overline{pq}\) can be written as \(u = tp + (1 - t)q,\) for some 0 ≤ t ≤ 1 and its corresponding point v on \(\overline{x_{k+1}m_{k}}\) is the point \(v = tx_{k+1} + (1 - t)m_{k}.\) Since the triangle Δ is unique up to a rigid transform, we can, without loss of generality, assume that the two triangles Δ, Σ are contained in a two-dimensional subspace of T m such that (using the obvious coordinates) they are spanned by the following two sets of three points:
with θ σ ≤ θ δ . Consequently, \(u = (ta + (1 - t)b\cos \theta _{\delta },(1 - t)b\sin \theta _{\delta })\) and \(v = (ta + (1 - t)b\cos \theta _{\sigma },(1 - t)b\sin \theta _{\sigma }),\) and their lengths are, respectively,
Since \(\theta _{\sigma } \leq \theta _{\delta }\), it then follows that \(\|u\| \leq \| v\|\).
Rights and permissions
Copyright information
© 2016 Springer International Publishing Switzerland
About this chapter
Cite this chapter
Cheng, G., Ho, J., Salehian, H., Vemuri, B.C. (2016). Recursive Computation of the Fréchet Mean on Non-positively Curved Riemannian Manifolds with Applications. In: Turaga, P., Srivastava, A. (eds) Riemannian Computing in Computer Vision. Springer, Cham. https://doi.org/10.1007/978-3-319-22957-7_2
Download citation
DOI: https://doi.org/10.1007/978-3-319-22957-7_2
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-22956-0
Online ISBN: 978-3-319-22957-7
eBook Packages: EngineeringEngineering (R0)