A new class of median based impulse rejecting filters

TLDR: A novel adaptive filter based on the impulse rejecting mechanism, where detected impulses are filtered and noise-free pixels are left unaltered, which consistently performs well in suppressing both types of impulse noise while still employing a simple structure.

Abstract: This paper proposes a novel adaptive filter based on the impulse rejecting mechanism, where detected impulses are filtered and noise-free pixels are left unaltered. Previous impulse detection strategies based on thresholding operations tend to work well for large, fixed-valued impulses but poorly for random-valued impulse noise, or vice versa. The objective of this work is to utilize the center weighted median (CWM) filters with variable center weights to define a more general operator, which forms estimates according to the differences defined between the outputs of CWM filters and the current pixel in consideration. As compared with existing schemes, the proposed filter consistently performs well in suppressing both types of impulse noise while still employing a simple structure. Better performance has been achieved by the new filter in restoring a variety of images corrupted with different noise ratios.