I am having an image consisting of {0,1,2} values and I am trying to run gaussian_filter from scipy.ndimage over it. Though it was very easy to use but I haven't understood the things properly yet.
Few things that I understood are :
- sigma is the standard deviation used for assigning weights in the kernel
- Gaussian filter on a 2D image is implemented in the form of 1D gaussian filters
- the size of the kernel depends upon sigma as well as truncate value [Gaussian filter in scipy
I want to know the weights being used in the gaussian filter. Is there some way to get the kernel used from the gaussian_filter call itself or some formula by which I can have that kernel.
I want this just for better understanding of weightage I am giving to different neighboring pixels in my image.
for example what will be kernel for
ndimage.gaussian_filter(image,sigma=0.2)
ndimage.gaussian_filter(image,sigma=0.2,truncate=2)