python - How to create a bidimensional Gaussian filter on a dense list of points -

i doing best replicate algorithm described here in paper making inpainting algorithm. idea contour or edge points of part of image needs inpainted. in order find linear point in region, orthogonal normal vector found. on page 6, short description of implementation given.

in our implementation contour δΩ of target region modelled dense list of image point locations. given point p ∈ δΩ, normal direction np computed follows: i) positions of “control” points of δΩ filtered via bi-dimensional gaussian kernel and, ii) np estimated unit vector orthogonal line through preceding , successive points in list.

so appears need put these points in gaussian filter. how set bi-dimensional gaussian filter when have single dimension or list of points?

lets our contour box shape @ points, create 1 dimensional list of points: [1,1],[1,2],[1,3],[2,1],[2,3],[3,1],[3,2],[3,3]. need make new 2d matrix table , put points in , leave middle point @ [2,2] empty, run gaussian filter on it? doesn't seem dense though.

i trying run through python libraries.

a dense list of image points

is line.

you applying gaussian filter black , white image line black , background white, understand. think doing that, approximate curve model fitting.

convolve of points in 2d region surrounding point , overwrite point result.

this make curve on edge of target region less sharp, lowering noise in calculation of normal, vector orthogonal 2 points surround current one.