We consider the role of scale in the context of the recently-developed non-local-means (NL-means) filter. A new example-based variant of the NL-means is introduced and results based on same-scale and cross-scale counterparts will be compared for a set of images. We consider the cases in which neighborhoods are taken from the observed image itself as well as from other irrelevant images, varying the smoothness parameter as well. Our experiments indicate that using cross-scale (i.e., downsampled) neighborhoods in the NL-means filter yields results that are quite comparable to those obtained by using neighborhoods at the same-scale. © 2008 Springer-Verlag Berlin Heidelberg.
CITATION STYLE
Ebrahimi, M., & Vrscay, E. R. (2008). Examining the role of scale in the context of the non-local-means filter. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 5112 LNCS, pp. 170–181). https://doi.org/10.1007/978-3-540-69812-8_17
Mendeley helps you to discover research relevant for your work.