Question about the sigma indexes inside FNLMD formula

asked 2018-06-11 16:06:24 -0600

rezaee gravatar image

updated 2018-06-17 17:13:12 -0600

I am interested to know, how a summation(Sigma) works if there are multi indexes under it? (Like what I have underlined by red in the following image)

EDIT(add more information):

The denoising algorithm implemented in OpenCV is called Fast Non-Local Means Denoising (FNLMD), and is based on work by Antoni Buades, Bartomeu Coll, and Jean-Michel Morel

The identification of similar pixels proceeds based on a window B(p, s) centered on pixel p and of size s. Given such a window around the point we wish to update, we can compare that window with an analogous window around some other pixel q. We define the square distance between B(p, s) and B(q, s) to be:

image description

Where c is the color index, I c (p) is the intensity of the image in channel c at point p, and the summation over j is over the elements of the patch. From this square dis‐ tance, a weight can be assigned to every other pixel relative to the pixel currently being updated.

edit retag flag offensive close merge delete


add some context please. where is that from ?

berak gravatar imageberak ( 2018-06-12 04:28:09 -0600 )edit

@berak: I found this formula in Oreally's Opencv3 book. I edited my question and added more information.

rezaee gravatar imagerezaee ( 2018-06-17 17:09:35 -0600 )edit