java - Directional Weighted Median Filter (Image processing) -


i need implement directional weighted median filter in java remove random impulse noise. have no idea how/where start. algorithm per below:

  1. create 5x5 window
  2. consider 4 directions (vertical, horizontal, diagonal left, diagonal right) center pixel (5 pixels in each direction)
  3. calculate weighted difference , take minimum value
  4. minimum value compared threshold value:
    if value > threshold: noise pixel
    else: not noise pixel
  5. calculate standard deviation of 5 pixels in each direction
  6. giving weight direction in standard deviation smallest, weighted median computed
  7. the noisy pixel replaced median value
  8. move window throughout image
  9. iterate steps 8 10 times

can point me in right direction how should go implementing this? examples or implemented codes highly appreciated. using imagej, plugin has implemented filter (or variation of it) helpful. thanks.

apparently, filter hasn't been implemented imagej plugin yet. unfortunately, couldn't find source in publication you've linked to. unless directly ask authors code, you'll have implement filter yourself, based on published description.

you can start implementations (and sources) of other filters listed @ imagej plugins page. i'd suggest implement extendedpluginfilter interface.

alternatively, should consider using imglib, generic image processing library included in fiji , being core model of future imagej2. can further on fiji mailing list (fiji@fiji.sc).


Comments

Popular posts from this blog

Perl - how to grep a block of text from a file -

delphi - How to remove all the grips on a coolbar if I have several coolbands? -

javascript - Animating array of divs; only the final element is modified -