On Learning Rotations

An algorithm is presented for online learning of rotations. The proposed algorithm involves matrix exponentiated gradient updates and is motivated by the von Neumann divergence. The multiplicative updates are exponentiated skew-symmetric matrices which comprise the Lie algebra of the rotation group. The orthonormality and unit determinant of the matrix parameter are preserved using matrix logarithms and exponentials and the algorithm lends itself to intuitive interpretation in terms of the differential geometry of the manifold associated with the rotation group. A complexity reduction result is presented that exploits the eigenstructure of the matrix updates to simplify matrix exponentiation to a quadratic form.

[1]  R. H. Oppermann,et al.  Introductory quantum mechanics , 1939 .

[2]  G. Wahba A Least Squares Estimate of Satellite Attitude , 1965 .

[3]  Grace Wahba,et al.  Problem 65-1: A least squares estimate of satellite attitude , 1966 .

[4]  P. Schönemann,et al.  A generalized solution of the orthogonal procrustes problem , 1966 .

[5]  J. Thorpe Elementary Topics in Differential Geometry , 1979 .

[6]  Paul J. Besl,et al.  A Method for Registration of 3-D Shapes , 1992, IEEE Trans. Pattern Anal. Mach. Intell..

[7]  D. Hestenes,et al.  Lie-groups as Spin groups. , 1993 .

[8]  Elijah Polak,et al.  Optimization: Algorithms and Consistent Approximations , 1997 .

[9]  Robert B. Fisher,et al.  Estimating 3-D rigid body transformations: a comparison of four major algorithms , 1997, Machine Vision and Applications.

[10]  Manfred K. Warmuth,et al.  Exponentiated Gradient Versus Gradient Descent for Linear Predictors , 1997, Inf. Comput..

[11]  L. Butler Applications of matrix theory to approximation theory , 1999 .

[12]  P. Evans Rotations and rotation matrices. , 2001, Acta crystallographica. Section D, Biological crystallography.

[13]  J. Gallier,et al.  COMPUTING EXPONENTIALS OF SKEW-SYMMETRIC MATRICES AND LOGARITHMS OF ORTHOGONAL MATRICES , 2002 .

[14]  Örjan Smedby,et al.  Compact and efficient 3D shape description through radial function approximation , 2003, Comput. Methods Programs Biomed..

[15]  Joan Lasenby,et al.  Applications of Conformal Geometric Algebra in Computer Vision and Graphics , 2004, IWMM/GIAE.

[16]  Gunnar Rätsch,et al.  Matrix Exponentiated Gradient Updates for On-line Learning and Bregman Projection , 2004, J. Mach. Learn. Res..

[17]  Kostas Daniilidis,et al.  Rotation recovery from spherical images without correspondences , 2006, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[18]  Kaare Brandt Petersen,et al.  The Matrix Cookbook , 2006 .

[19]  Thierry Paul,et al.  Quantum computation and quantum information , 2007, Mathematical Structures in Computer Science.

[20]  Manfred K. Warmuth Winnowing subspaces , 2007, ICML '07.

[21]  Manfred K. Warmuth,et al.  Learning Rotations , 2008, COLT.

[22]  Visa Koivunen,et al.  Steepest Descent Algorithms for Optimization Under Unitary Matrix Constraint , 2008, IEEE Transactions on Signal Processing.

[23]  Raman Arora,et al.  Navigation using a spherical camera , 2008, 2008 19th International Conference on Pattern Recognition.

[24]  W. Sethares,et al.  Group theoretical methods in signal processing: learning similarities, transformations and invariants , 2009 .

[25]  Pascal Frossard,et al.  3D Face Recognition with Sparse Spherical Representations , 2008, Pattern Recognit..