• Corpus ID: 31022153

A multigrid framework for variational approaches in medical image processing and computer vision

  title={A multigrid framework for variational approaches in medical image processing and computer vision},
  author={Harald K{\"o}stler},
Variational approaches for modeling are very common in medical image processing and computer vision, e.g., for motion detection by optical flow, image registration, image segmentation, image denoising, or image inpainting. One drawback of variational approaches is that the optimization process typically involves the solution of one or more partial differential equations (PDEs). This is very time and memory consuming, since the amount of data that has to be processed in imaging increases more… 
Performance engineering to achieve real-time high dynamic range imaging
A rigorous performance analysis is done and a performance model for the multigrid algorithm is derived that guides us to an improved implementation, where the overall performance of more than 25 frames per second for 16.8 Megapixel images doing full high dynamic range compression including data transfers between CPU and GPU.
A Hybrid Iterative Algorithm For Reconstruction Of X-Ray Computed Tomography
A hybrid iterative algorithm by combining multigrid method, Tikhonov regularization and Simultaneous Iterative Reconstruction Technique for reconstruction of the computed tomography image that reduces the time and the volume of computations considerably.
Modeling Multigrid Algorithms for Variational Imaging
This work discusses the extension of Syntony to include automatic code generation in the context of large scale continuous simulations that require the numerical solution of partial differential equations (PDE) and evaluates the applicability of the approach in a case study for image denoising.
Extension of a Software Package for Hierarchical Hybrid Grids
The intention is not just to measure the influence of the kernels on the performance, but also to get a better understanding of speed limiting factors, therefore performance modeling is applied and discussed.
Quantitative material decomposition using linear iterative near-field phase retrieval dual-energy x-ray imaging
The LIPR formalism is expanded to achieve quantitative material thickness decomposition and is presented the first application of dual-energy phase retrieval with the AM model to monochromatic experimental X-rays projections at two different energy for obtaining split X-ray interactions.
Lehrstuhl für Informatik 10 ( Systemsimulation ) On the Construction of Prolongation Operators in Multigrid
This work discusses a general framework for the construction of prolongation operators for multigrid methods and shows that classical matrix-dependent transfers like black box prolongation can be recovered as special cases and presents numerical results for several two and three-dimensional diffusion problems including jumping coefficients.
Lehrstuhl für Informatik 10 ( Systemsimulation ) Multigrid HowTo : A simple Multigrid solver in C + + in less than 200 lines of code
A 2D Poisson solver on a structured cell-based grid that requires less than 200 lines of code is shown and a step-by-step introduction to multigrid from a practical point of view is given.
Lehrstuhl für Informatik 10 ( Systemsimulation ) Multigrid HowTo : An Open Source Algebraic Multigrid Solver in C + +
This report is a continuation of the technical report 08-3 A simple Multigrid solver in C++ in less than 200 lines of code and presents some simple test results.
Une approche basée sur les modèles pour le développement d'applications de simulation numérique haute-performance. (A Model-Based Approach for the Development of High-Performance Scientific Computing Software)
Cette approche nommee MDE4HPC, definit un langage de modelisation dedie permet aux numericiens de decrire la resolution de leurs modeles theoriques dans un langages qui d'une part leur est familier and d'autre part est independant d'un quelconque architecture materielle.