Ranked-Based Vector Median Filter
Bogdan Smołka · Hokkaido University Collection of Scholarly and Academic Papers (Hokkaido University) · 2009
In this paper a new class of noise attenuating and edge enhancing filters for color image processing is proposed. The new filtering design is based, like the vector median filter, on the minimization of aggregated weighted distances between pixels in a filtering window. The weights which are assigned to the sequence of the sorted distances are decreasing functions of their ranks. In this way, the largest distances are taken to the aggregated distance with smallest weight, which ensures much better noise reduction properties of the described filter. Additionally, the proposed method has a unique ability to strengthen the edges of the noisy image objects. This unique deblurring property can be useful in many applications which require reliable impulse noise reduction and efficient edge sharpening of color images.