Is there any limit to the region sampled and averaged. I think I understand how the gaussian kernal is used to weight the nearer pixels, but it would seem like at a certain point it would make sense to only sample pixels within a certain distance of the current pixel, instead of the whole image.
moridin22
Yep, that's the idea of the "support region of the kernel" in this slide - this basically means the region where the kernel is nonzero. In the previous examples, the kernel was given by a matrix of some size, where the implication was that outside of that matrix, the kernel was just zero.
Is there any limit to the region sampled and averaged. I think I understand how the gaussian kernal is used to weight the nearer pixels, but it would seem like at a certain point it would make sense to only sample pixels within a certain distance of the current pixel, instead of the whole image.
Yep, that's the idea of the "support region of the kernel" in this slide - this basically means the region where the kernel is nonzero. In the previous examples, the kernel was given by a matrix of some size, where the implication was that outside of that matrix, the kernel was just zero.