Measure-Valued Variational Models with Applications to Diffusion-Weighted Imaging

We develop a general mathematical framework for variational problems where the unknown function takes values in the space of probability measures on some metric space. We study weak and strong topologies and define a total variation seminorm for functions taking values in a Banach space. The seminorm penalizes jumps and is rotationally invariant under certain conditions. We prove existence of a minimizer for a class of variational problems based on this formulation of total variation and provide an example where uniqueness fails to hold. Employing the Kantorovich–Rubinstein transport norm from the theory of optimal transport, we propose a variational approach for the restoration of orientation distribution function-valued images, as commonly used in diffusion MRI. We demonstrate that the approach is numerically feasible on several data sets.

[1]  Antje Baer,et al.  Direct Methods In The Calculus Of Variations , 2016 .

[2]  Ying Wu,et al.  Vectorial total variation regularisation of orientation distribution functions in diffusion weighted MRI , 2014, Int. J. Bioinform. Res. Appl..

[3]  Luigi Ambrosio,et al.  Metric space valued functions of bounded variation , 1990 .

[4]  Min Li,et al.  Adaptive Primal-Dual Splitting Methods for Statistical Learning and Image Processing , 2015, NIPS.

[5]  Marco Cuturi,et al.  Sinkhorn Distances: Lightspeed Computation of Optimal Transport , 2013, NIPS.

[6]  Alan Connelly,et al.  Robust determination of the fibre orientation distribution in diffusion MRI: Non-negativity constrained super-resolved spherical deconvolution , 2007, NeuroImage.

[7]  Edwin Hewitt,et al.  Real And Abstract Analysis , 1967 .

[8]  Remco Duits,et al.  Left-Invariant Diffusions on the Space of Positions and Orientations and their Application to Crossing-Preserving Smoothing of HARDI images , 2011, International Journal of Computer Vision.

[9]  Rachid Deriche,et al.  Quantitative Comparison of Reconstruction Methods for Intra-Voxel Fiber Recovery From Diffusion MRI , 2014, IEEE Transactions on Medical Imaging.

[10]  Thorsten Hohage,et al.  A Coherence Enhancing Penalty for Diffusion MRI: Regularizing Property and Discrete Approximation , 2014, SIAM J. Imaging Sci..

[11]  Anuj Srivastava,et al.  Riemannian Analysis of Probability Density Functions with Applications in Vision , 2007, 2007 IEEE Conference on Computer Vision and Pattern Recognition.

[12]  Max A. Viergever,et al.  Recursive calibration of the fiber response function for spherical deconvolution of diffusion MRI data , 2014, NeuroImage.

[13]  Antonin Chambolle,et al.  A First-Order Primal-Dual Algorithm for Convex Problems with Applications to Imaging , 2011, Journal of Mathematical Imaging and Vision.

[14]  Remco Duits,et al.  Numerical Schemes for Linear and Non-linear Enhancement of DW-MRI , 2011, SSVM.

[15]  Fernando Pérez,et al.  Sparse reproducing kernels for modeling fiber crossings in diffusion weighted imaging , 2013, 2013 IEEE 10th International Symposium on Biomedical Imaging.

[16]  Hugo Lavenant.  Harmonic mappings valued in the Wasserstein space , 2017, Journal of Functional Analysis.

[17]  Alan Connelly,et al.  Direct estimation of the fiber orientation density function from diffusion-weighted MRI data using spherical deconvolution , 2004, NeuroImage.

[18]  Miloslav Duchoˇn,et al.  Functions with bounded variation in locally convex space , 2011 .

[19]  F. Clarke.  Functional Analysis, Calculus of Variations and Optimal Control , 2013 .

[20]  B. Dacorogna.  Direct methods in the calculus of variations , 1989 .

[21]  R. Duits,et al.  New Exact and Numerical Solutions of the (Convection-)Diffusion Kernels on SE(3) , 2016, 1604.03843.

[22]  Xiaoming Yuan,et al.  Adaptive Primal-Dual Hybrid Gradient Methods for Saddle-Point Problems , 2013, 1305.0546.

[23]  Maxime Descoteaux,et al.  Contextual Diffusion Image Post-processing Aids Clinical Applications , 2015, Visualization and Processing of Higher Order Descriptors for Multi-Valued Data.

[24]  Freddie Åström,et al.  Image Labeling by Assignment , 2016, Journal of Mathematical Imaging and Vision.

[25]  Kaleem Siddiqi,et al.  3D Stochastic Completion Fields for Mapping Connectivity in Diffusion MRI , 2013, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[26]  C. Ionescu Tulcea,et al.  Topics in the Theory of Lifting , 1969 .

[27]  Carola-Bibiane Schönlieb,et al.  Imaging with Kantorovich-Rubinstein Discrepancy , 2014, SIAM J. Imaging Sci..

[28]  Yogesh Rathi,et al.  On Approximation of Orientation Distributions by Means of Spherical Ridgelets , 2008, IEEE Transactions on Image Processing.

[29]  M. Slemrod,et al.  PDEs and continuum models of phase transitions : proceedings of an NSF-CNRS joint seminar held in Nice, France, January 18-22, 1988 , 1989 .

[30]  Michael Möller,et al.  Sublabel–Accurate Relaxation of Nonconvex Energies , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[31]  H. Pfeifer.  Principles of Nuclear Magnetic Resonance Microscopy , 1992 .

[32]  Antonin Chambolle,et al.  Total roto-translational variation , 2017, Numerische Mathematik.

[33]  Baba C. Vemuri,et al.  Variational denoising of diffusion weighted MRI , 2009 .

[34]  Daniel Cremers,et al.  The Natural Vectorial Total Variation Which Arises from Geometric Measure Theory , 2012, SIAM J. Imaging Sci..

[35]  Marco Reisert,et al.  Spherical Tensor Algebra: A Toolkit for 3D Image Processing , 2017, Journal of Mathematical Imaging and Vision.

[36]  Michele Miranda,et al.  Functions of bounded variation on “good” metric spaces , 2003 .

[37]  WeinmannAndreas,et al.  Mumford---Shah and Potts Regularization for Manifold-Valued Data , 2016 .

[38]  Marco Reisert,et al.  Fiber Continuity Based Spherical Deconvolution in Spherical Harmonic Domain , 2013, MICCAI.

[39]  Yann Gousseau,et al.  The TVL1 Model: A Geometric Point of View , 2009, Multiscale Model. Simul..

[40]  Ron Wakkary,et al.  Integration , 2016, Interactions.

[41]  Anuj Srivastava,et al.  A novel Riemannian metric for analyzing HARDI data , 2011, Medical Imaging.

[42]  Frithjof Kruggel,et al.  A Reproducing Kernel Hilbert Space Approach for Q-Ball Imaging , 2011, IEEE Transactions on Medical Imaging.

[43]  U. Klose,et al.  Regularization of bending and crossing white matter fibers in MRI Q-ball fields. , 2011, Magnetic resonance imaging.

[44]  Jan Lellmann,et al.  An Optimal Transport-Based Restoration Method for Q-Ball Imaging , 2017, SSVM.

[45]  Massimo Fornasier,et al.  Theoretical Foundations and Numerical Methods for Sparse Recovery , 2010, Radon Series on Computational and Applied Mathematics.

[46]  D. Tuch.  Q‐ball imaging , 2004, Magnetic resonance in medicine.

[47]  J. E. Tanner,et al.  Spin diffusion measurements : spin echoes in the presence of a time-dependent field gradient , 1965 .

[48]  A. Chambolle,et al.  An introduction to Total Variation for Image Analysis , 2009 .

[49]  Maxime Descoteaux,et al.  Dipy, a library for the analysis of diffusion MRI data , 2014, Front. Neuroinform..

[50]  Ignace Lemahieu,et al.  POSTPROCESSING OF BRAIN WHITE MATTER FIBER ORIENTATION DISTRIBUTION FUNCTIONS , 2007, 2007 4th IEEE International Symposium on Biomedical Imaging: From Nano to Macro.

[51]  Gabriele Steidl,et al.  Transport Between RGB Images Motivated by Dynamic Optimal Transport , 2015, Journal of Mathematical Imaging and Vision.

[52]  P. Basser,et al.  MR diffusion tensor spectroscopy and imaging. , 1994, Biophysical journal.

[53]  Elias Kellner,et al.  About the Geometry of Asymmetric Fiber Orientation Distributions , 2012, IEEE Transactions on Medical Imaging.

[54]  Andreas Weinmann,et al.  Mumford–Shah and Potts Regularization for Manifold-Valued Data , 2014, Journal of Mathematical Imaging and Vision.

[55]  Stamatios N. Sotiropoulos,et al.  Spherical Deconvolution of Multichannel Diffusion MRI Data with Non-Gaussian Noise Models and Spatial Regularization , 2014, PloS one.

[56]  M. Descoteaux.  High angular resolution diffusion MRI : from local estimation to segmentation and tractography , 2008 .

[57]  John M. Lee.  Riemannian Manifolds: An Introduction to Curvature , 1997 .

[58]  Daniel Cremers,et al.  Total Variation Regularization for Functions with Values in a Manifold , 2013, 2013 IEEE International Conference on Computer Vision.

[59]  E. Özarslan,et al.  Asymmetric Orientation Distribution Functions (AODFs) revealing intravoxel geometry in diffusion MRI. , 2018, Magnetic resonance imaging.

[60]  Yunho Kim,et al.  HARDI DATA DENOISING USING VECTORIAL TOTAL VARIATION AND LOGARITHMIC BARRIER. , 2010, Inverse problems and imaging.

[61]  N. Makris,et al.  High angular resolution diffusion imaging reveals intravoxel white matter fiber heterogeneity , 2002, Magnetic resonance in medicine.

[62]  A. Mondino.  ON RIEMANNIAN MANIFOLDS , 1999 .

[63]  Alfred Anwander,et al.  Position-orientation adaptive smoothing of diffusion weighted magnetic resonance data (POAS) , 2012, Medical Image Anal..

[64]  Tony F. Chan,et al.  Aspects of Total Variation Regularized L[sup 1] Function Approximation , 2005, SIAM J. Appl. Math..

[65]  Christophe Lenglet,et al.  Estimating Orientation Distribution Functions with Probability Density Constraints and Spatial Regularity , 2009, MICCAI.

[66]  Kellen Petersen August.  Real Analysis , 2009 .

[67]  T. Goldstein.  Adaptive Primal Dual Optimization for Image Processing and Learning , 2013 .

[68]  C. Villani.  Optimal Transport: Old and New , 2008 .

[69]  Antonin Chambolle,et al.  Diagonal preconditioning for first order primal-dual algorithms in convex optimization , 2011, 2011 International Conference on Computer Vision.

[70]  L. Ambrosio,et al.  Functions of Bounded Variation and Free Discontinuity Problems , 2000 .

[71]  Laurent D. Cohen,et al.  Global Minimum for a Finsler Elastica Minimal Path Approach , 2016, International Journal of Computer Vision.

[72]  Remco Duits,et al.  Improving Fiber Alignment in HARDI by Combining Contextual PDE Flow with Constrained Spherical Deconvolution , 2015, PloS one.

[73]  Christophe Lenglet,et al.  ODF reconstruction in q-ball imaging with solid angle consideration , 2009, 2009 IEEE International Symposium on Biomedical Imaging: From Nano to Macro.

[74]  Michael Möller,et al.  Sublabel-Accurate Convex Relaxation of Vectorial Multilabel Energies , 2016, ECCV.

[75]  Viola Priesemann,et al.  Local active information storage as a tool to understand distributed neural information processing , 2013, Front. Neuroinform..

[76]  J. Ball.  A version of the fundamental theorem for young measures , 1989 .

[77]  R. Duits,et al.  Morphological and Linear Scale Spaces for Fiber Enhancement in DW-MRI , 2013, J. Math. Imaging Vis..

[78]  Gabriele Steidl,et al.  A Second Order Nonsmooth Variational Model for Restoring Manifold-Valued Images , 2015, SIAM J. Sci. Comput..

[79]  Bernhard Schmitzer,et al.  Optimal Transport for Manifold-Valued Images , 2017, SSVM.

[80]  Remco Duits,et al.  Fast implementations of contextual PDE's for HARDI data processing in DIPY , 2016 .

[81]  P Ossenblok,et al.  Cleaning output of tractography via fiber to bundle coherence, a new open source implementation , 2016 .