iPAUR

(image Processing At University of Regensburg)

The code is implemented for the master thesis found in the src/LaTex folder.

Main Publications

A list of the main publications we follow in this thesis and coding:

Chambolle Antonin and Pock Thomas. A first-order primal-dual algorithm for convex problems with applications to imaging. ”Journal of Mathematical Imaging and Vision”, 40(1):120–145, 2011.

A. Chambolle, V. Caselles, D. Cremers, M. Novaga, and T. Pock. An introduction to total variation for image analysis. In ”Theoretical Foundations and Numerical Methods for Sparse Recovery”. De Gruyter, 2010.

T. Pock, D. Cremers, H. Bischof, and A. Chambolle. An algorithm for minimizing the piecewise smooth mumford-shah functional. In ”IEEE International Conference on Computer Vision (ICCV)”, Kyoto, Japan, 2009.

E. Strekalovskiy and D. Cremers. Real-time minimization of the piecewise smooth mumford-shah functional. In ”European Conference on Computer Vision (ECCV)”, pages 127–141, 2014. https://github.com/tum-vision/fastms.

Image References

• img/examples is taken from https://github.com/tum-vision/fastms.git and belongs to the last publication in the list above
• img/lena is the well known Lena test image
• img/pock are images taken from the third publication of the above list
• img/keating is a scan of a brain tumor provided by Steven Keating, see for instance http://stevenkeating.info/main.html

Implemented Algorithms and Basic Image Processing

• Serial algorithms:
  • ROF-Model
  • TVL1-Model
  • Image Inpainting
  • Huber-ROF-Model
  • Real-Time Minimizer for the Mumford-Shah functional
• Other basic serial algorithms:
  • mean value blur
  • gaussian blur
  • canny edge detection
  • dilatation
  • erosion
  • duto filter
  • gradient filter (Sobel, Prewitt, Robert's Cross)
  • color space conversions
  • inverse image
  • laplace filter (operator)
  • median filter
• Parallel algorithms:
  • Primal-Dual Algorithm to solve the convex relaxed Mumford-Shah functional using Dykstra's projection algorithm
  • Primal-Dual Algorithm to solve the convex relaxed Mumford-Shah functional using an approach with Lagrange multipliers
  • ROF-Model
  • TVL1-Model
  • Image Inpainting
  • Real-Time Minimizer for the Mumford-Shah functional

Requirements

CUDA

For the use of the GPU implementation, CUDA is needed.

OpenCV

For reading and writing images OpenCV is needed.