Image restoration based on Laplacian preprocessed long-range correlation

  • Authors:
  • Javad Ahmadi-Shokouh;Hengameh Keshavarz

  • Affiliations:
  • Department of Electrical & Computer Engineering, University of Waterloo, Waterloo, Canada N2L 3G1;Department of Electrical & Computer Engineering, University of Waterloo, Waterloo, Canada N2L 3G1

  • Venue:
  • Multidimensional Systems and Signal Processing
  • Year:
  • 2008

Quantified Score

Hi-index 0.00

Visualization

Abstract

A decision-based adaptive filter is proposed to remove impulse-noise: (1) A modified Laplacian operator accurately locates impulse-noise corrupted pixels. (2) Another image-region highly correlated with the to-be-restored image-region is identified through "Long-Range Correlation" with statistically designed search-ranges and match-criterion. (3) The center of the to-be-restored image-region is replaced using a proposed recovering approach. Simulation results demonstrate the proposed method's superior impulse-noise removal capability vis-a-vis other recent approaches.