Regularizing Flows for Constrained Matrix-Valued Images

Nonlinear diffusion equations are now widely used to restore and enhance images. They allow to eliminate noise and artifacts while preserving large global features, such as object contours. In this context, we propose a differential-geometric framework to define PDEs acting on some manifold constrained datasets. We consider the case of images taking value into matrix manifolds defined by orthogonal and spectral constraints. We directly incorporate the geometry and natural metric of the underlying configuration space (viewed as a Lie group or a homogeneous space) in the design of the corresponding flows. Our numerical implementation relies on structure-preserving integrators that respect intrinsically the constraints geometry. The efficiency and versatility of this approach are illustrated through the anisotropic smoothing of diffusion tensor volumes in medical imaging.

[1]  W. Boothby.  An introduction to differentiable manifolds and Riemannian geometry , 1975 .

[2]  S. Helgason.  Differential Geometry, Lie Groups, and Symmetric Spaces , 1978 .

[3]  J. Marsden,et al.  Product formulas and numerical algorithms , 1978 .

[4]  C. Botsaris.  Constrained optimization along geodesics , 1981 .

[5]  Charles R. Johnson,et al.  Matrix analysis , 1985, Statistical Inference for Engineers and Data Scientists.

[6]  F. Testard,et al.  Introduction a la théorie des groupes de Lie classiques , 1986 .

[7]  R. Abraham,et al.  Manifolds, tensor analysis, and applications: 2nd edition , 1988 .

[8]  Karen K. Uhlenbeck.  Harmonic maps into Lie groups: classical solutions of the chiral model , 1989 .

[9]  P. Crouch,et al.  Numerical integration of ordinary differential equations on manifolds , 1993 .

[10]  H. Urakawa.  Calculus of variations and harmonic maps , 1993 .

[11]  U. Helmke,et al.  Optimization and Dynamical Systems , 1994, Proceedings of the IEEE.

[12]  H. Urakawa,et al.  Harmonic maps into Lie groups and homogeneous spaces , 1997 .

[13]  Pietro Perona.  Orientation diffusions , 1998, IEEE Trans. Image Process..

[14]  Ron Kimmel,et al.  A general framework for low level vision , 1998, IEEE Trans. Image Process..

[15]  Alan Edelman,et al.  The Geometry of Algorithms with Orthogonality Constraints , 1998, SIAM J. Matrix Anal. Appl..

[16]  Shun-ichi Amari,et al.  Natural Gradient Works Efficiently in Learning , 1998, Neural Computation.

[17]  C. Poupon.  Detection des faisceaux de fibres de la substance blanche pour l'etude de la connectivite anatomique cerebrale , 1999 .

[18]  A. Iserles,et al.  Lie-group methods , 2000, Acta Numerica.

[19]  Simon R. Arridge,et al.  A Regularization Scheme for Diffusion Tensor Magnetic Resonance Images , 2001, IPMI.

[20]  Rachid Deriche,et al.  Diffusion tensor regularization with constraints preservation , 2001, Proceedings of the 2001 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. CVPR 2001.

[21]  B. Vemuri,et al.  Fiber tract mapping from diffusion tensor MRI , 2001, Proceedings IEEE Workshop on Variational and Level Set Methods in Computer Vision.

[22]  G. Sapiro,et al.  Geometric partial differential equations and image analysis [Book Reviews] , 2001, IEEE Transactions on Medical Imaging.

[23]  Gjlles Aubert,et al.  Mathematical problems in image processing , 2001 .

[24]  Tony F. Chan,et al.  Variational Restoration of Nonflat Image Features: Models and Algorithms , 2001, SIAM J. Appl. Math..

[25]  S. Osher,et al.  Variational problems and PDEs on implicit surfaces , 2001, Proceedings IEEE Workshop on Variational and Level Set Methods in Computer Vision.

[26]  S. Osher,et al.  Solving variational problems and partial differential equations mapping into general target manifolds , 2004 .

[27]  Carl-Fredrik Westin,et al.  Processing and visualization for diffusion tensor MRI , 2002, Medical Image Anal..

[28]  Stanley Osher,et al.  Numerical Methods for p-Harmonic Flows and Applications to Image Processing , 2002, SIAM J. Numer. Anal..

[29]  Rachid Deriche,et al.  Constrained Flows of Matrix-Valued Functions: Application to Diffusion Tensor Regularization , 2002, ECCV.

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

[31]  T. Brox,et al.  Diffusion and regularization of vector- and matrix-valued images , 2002 .

[32]  Peter J. Olver,et al.  Geometric Integration Algorithms on Homogeneous Manifolds , 2002, Found. Comput. Math..

[33]  Ron Kimmel,et al.  Orientation Diffusion or How to Comb a Porcupine , 2002, J. Vis. Commun. Image Represent..

[34]  Ronald Fedkiw,et al.  Level set methods and dynamic implicit surfaces , 2002, Applied mathematical sciences.

[35]  E. Hairer,et al.  Geometric Numerical Integration: Structure Preserving Algorithms for Ordinary Differential Equations , 2004 .

[36]  Alain Trouvé,et al.  Diffeomorphisms Groups and Pattern Matching in Image Analysis , 1998, International Journal of Computer Vision.

[37]  Guillermo Sapiro,et al.  Diffusion of General Data on Non-Flat Manifolds via Harmonic Maps Theory: The Direction Diffusion Case , 2000, International Journal of Computer Vision.

[38]  Rachid Deriche,et al.  Orthonormal Vector Sets Regularization with PDE's and Applications , 2002, International Journal of Computer Vision.