High performance cluster computing with 3-D nonlinear diffusion filters

Abstract This paper deals with parallelization and implementation aspects of partial differential equation (PDE)-based image processing models for large cluster environments with distributed memory. As an example we focus on nonlinear diffusion filtering which we discretize by means of an additive operator splitting (AOS). We start by decomposing the algorithm into small modules that shall be parallelized separately. For this purpose image partitioning strategies are discussed and their impact on the communication pattern and volume is analyzed. Based on the results we develop an algorithmic implementation with excellent scaling properties on massively connected low-latency networks. Test runs on two different high-end Myrinet clusters yield almost linear speedup factors up to 209 for 256 processors. This results in typical denoising times of 0.4 s for five iterations on a 256×256×128 data cube.

[1]  Joachim Weickert,et al.  Efficient image segmentation using partial differential equations and morphology , 2001, Pattern Recognit..

[2]  Jitendra Malik,et al.  Scale-Space and Edge Detection Using Anisotropic Diffusion , 1990, IEEE Trans. Pattern Anal. Mach. Intell..

[3]  G. Koepfler,et al.  A morphological scheme for mean curvature motion and applications to anisotropic diffusion and motion of level sets , 1995 .

[4]  Max A. Viergever,et al.  Efficient and reliable schemes for nonlinear diffusion filtering , 1998, IEEE Trans. Image Process..

[5]  Michel Barlaud,et al.  Deterministic edge-preserving regularization in computed imaging , 1997, IEEE Trans. Image Process..

[6]  Timo Kohlberger,et al.  Designing 3-D Nonlinear Diffusion Filters for High Performance Cluster Computing , 2002, DAGM-Symposium.

[7]  Martin Rumpf,et al.  Nonlinear Diffusion in Graphics Hardware , 2001, VisSym.

[8]  Peng Lin,et al.  Lattice Boltzmann Models for Anisotropic Diffusion of Images , 2004, Journal of Mathematical Imaging and Vision.

[9]  Wolfgang Effelsberg,et al.  Fast Implicit Active Contour Models , 2002, DAGM-Symposium.

[10]  Christoph Schnörr,et al.  Globally convergent iterative numerical schemes for nonlinear variational image smoothing and segmentation on a multiprocessor machine , 2001, IEEE Trans. Image Process..

[11]  P. Lions,et al.  Image selective smoothing and edge detection by nonlinear diffusion. II , 1992 .

[12]  J. Weickert.  Applications of nonlinear diffusion in image processing and computer vision , 2000 .

[13]  Roman Goldenberg,et al.  Fast Geodesic Active Contours , 1999, Scale-Space.

[14]  Joachim Weickert,et al.  Anisotropic diffusion in image processing , 1996 .

[15]  Christoph Schnörr,et al.  A 1D Analog VLSI Implementation For Non-linear Real-Time Signal Preprocessing , 2001 .

[16]  J. Waldvogel,et al.  Numerical Analysis: A Comprehensive Introduction , 1989 .

[17]  Fred L. Fontaine,et al.  Wavelet‐based solution to anisotropic diffusion equation for edge detection , 1998 .

[18]  N. Paragios,et al.  Gradient vector flow fast geodesic active contours , 2001, Proceedings Eighth IEEE International Conference on Computer Vision. ICCV 2001.

[19]  H. De Man,et al.  A VLSI-architecture for parallel non-linear diffusion with applications in vision , 1994, Proceedings of 1994 IEEE Workshop on VLSI Signal Processing.

[20]  J. Weickert,et al.  Image Processing Using a Wavelet Algorithm for Nonlinear Diffusion , 1994 .

[21]  Christoph Schnörr,et al.  A One-Dimensional Analog VLSI Implementation for Nonlinear Real-Time Signal Preprocessing , 2001, Real Time Imaging.

[22]  Joachim Weickert,et al.  Scale-Space Theories in Computer Vision , 1999, Lecture Notes in Computer Science.

[23]  Martin Rumpf,et al.  An Adaptive Finite Element Method for Large Scale Image Processing , 1999, J. Vis. Commun. Image Represent..

[24]  Christoph Schnörr,et al.  Investigation of parallel and globally convergent iterative schemes for nonlinear variational image smoothing and segmentation , 1998, Proceedings 1998 International Conference on Image Processing. ICIP98 (Cat. No.98CB36269).

[25]  Joachim Weickert,et al.  Parallel implementations of AOS schemes: a fast way of nonlinear diffusion filtering , 1997, Proceedings of International Conference on Image Processing.

[26]  R. Temam.  Sur l'approximation de la solution des équations de Navier-Stokes par la méthode des pas fractionnaires (I) , 1969 .

[27]  R. Temam.  Sur l'approximation de la solution des équations de Navier-Stokes par la méthode des pas fractionnaires (II) , 1969 .

[28]  David S. Ebert,et al.  Data visualization 2001 : Proceedings of the joint Eurographics - IEEE TCVG Symposium on Visualization in Ascona, Switzerland, May 28-30, 2001 , 2001 .

[29]  Scott T. Acton,et al.  Multigrid anisotropic diffusion , 1998, IEEE Trans. Image Process..

[30]  Jitendra Malik,et al.  A network for multiscale image segmentation , 1988, 1988., IEEE International Symposium on Circuits and Systems.

[31]  K. R. Ramakrishnan,et al.  A Stochastic Scale Space for Multiscale Image Representation , 1999, Scale-Space.

[32]  Ravi Malladi,et al.  Fast Difference Schemes for Edge Enhancing Beltrami Flow , 2002, ECCV.

[33]  Joachim Weickert,et al.  Variational Optic Flow Computation with a Spatio-Temporal Smoothness Constraint , 2001, Journal of Mathematical Imaging and Vision.

[34]  J. Weickert,et al.  Fast Methods for Implicit Active Contour Models , 2003 .

[35]  Christoph Schnörr,et al.  A Study of a Convex Variational Diffusion Approach for Image Segmentation and Feature Extraction , 1998, Journal of Mathematical Imaging and Vision.

[36]  K. Mikula,et al.  A coarsening finite element strategy in image selective smoothing , 1997 .

[37]  Xuecheng Tai,et al.  A parallel splitting up method and its application to Navier-Stokes equations , 1991 .

[38]  J. Kacur,et al.  Solution of nonlinear diffusion appearing in image smoothing and edge detection , 1995 .

[39]  Fiorella Sgallari,et al.  Variational Numerical Methods for Solving Nonlinear Diffusion Equations Arising in Image Processing , 2002, J. Vis. Commun. Image Represent..

[40]  S. Osher,et al.  Geometric Level Set Methods in Imaging, Vision, and Graphics , 2011, Springer New York.