I want to know the kernel size applied to a Gaussian filter

You want to apply a Gaussian filter with a standard deviation of 2 to an image.
imgaussfilt(A, 2)
This image is 0.056 mm per pixel.
The blur is applied over a range of 2x0.056 mm, so I expect the boundaries to be softened within a range of 0.112 mm.
However, I was told that I should also consider the kernel size.
Is there any way to know the kernel size?

이미지에 표준편차가 2인 가우시안 필터를 적용하려고 해요.
imgaussfilt(A, 2)
이 이미지는 1pixel 당 0.056mm입니다.
블러가 적용되는 범위가 2x0.056mm로, 0.112mm 범위 내에서 경계가 부드러워질 것으로 예상했어요.
그런데, 커널 사이즈도 고려해야 한다고 하더라고요.
혹시 커널 사이즈를 알 수 있는 방법이 있을까요?

Answer 1

[Angelo Yeo]

In the document of imgaussfilt, it is mentioned that the default filter size is determined by the formula below:

2*ceil(2*sigma)+1

Because you set sigma of 2, the filter size will be 9.

sigma = 2;
2*ceil(2*sigma)+1

ans = 9