Ask Your Question

Revision history [back]

click to hide/show revision 1
initial version

pyramid_reduce, about sigma (down sampling)

Hi, I've studied SIFT and I reach here.

I know that gaussian blurring precedes down-sampling because of anti-aliasing.

from this link, 'https://scikit-image.org/docs/0.7.0/api/skimage.transform.pyramids.html#pyramid-reduce' , please see pyramid_reduce.

and the explanation is that

'sigma : float, optional Sigma for the gaussian filter. Default is 2 * downscale / 6.0 which corresponds to a filter mask twice the size of the scale factor that covers more than 99% of the gaussian distribution.'

I don't understand this sentence.

I know that 6sigma range includes 99% region of Gaussian distribution so, kernel size should be 6isgma. ex) sigma = 1, kernel size = 6sigma = 6 but, should be odd value, finally kernel size = 7

Can you explain to me the reason?