In this paper a novel method of impulsive noise removal in color images is presented. The proposed ltering designis based on a new measure of pixel similarity, which takes into account the structure of the local neighborhood ofthe pixels being compared. Thus, the new distance measure can be regarded as an extension of the reachabilitydistance used in the construction of the local outlier factor, widely used in the big data analysis. Using the newsimilarity measure, an extension of the classic Vector Median Filter (VMF) has been developed. The new lteris extremely robust to outliers introduced by the impulsive noise, retains details and has the unique ability tosharpen image edges. Using the structure of the developed lter, a new impulse detector has been constructed.The cumulated sum of smallest reachability distances in the ltering window serves as a robust measure of pixeloutlyingness. In this way, a pixel will be treated as corrupted if a prede ned threshold is exceeded and willbe replaced by the average of pixels which were found to belong to the original, pristine image; otherwise theprocessed pixel will be retained. This structure is similar to the Fast Averaging Peer Group Filter, however theincorporation of the reachability measure makes this technique more robust. The new ltering design can beapplied in real time scenario, as its computational e ciency is comparable with the standard VMF, which is fastenough to be used for the enhancement of video sequences. The new lter operates in a 3 3 ltering window,however the information acquired from a larger window is processed. The source of additional information isthe local neighborhood of pixels, which is used for the determination of the novel reachability measure. Theexperiments performed on a large database of color images show that the new lter surpasses existing designsespecially in the case of highly polluted images. The robust reachability measure assures that the clusters ofimpulses are being removed, as not only the pixels, but also their neighborhoods are considered. The novelmeasure of dissimilarity can be also used in other tasks whose main goal is the detection of outliers.
展开▼
