How to find corresponding values in a 3D matrix (2D images stacked as 3D) - Gaussian scale space.

Hello matlab community,I have a 2D image, which i have blurred continuously and stacked them over to form a 3D matrix. If "Gss" is my 3D matrix of stacked images - Gss(:,:,1) will be my original image with zero blur, Gss(:,:,2) will be a little blurred compared to the original image, Gss(:,:,3) will be even more blurred. So my task is to connect the critical points at one level, with its corresponding critical points at the next level. For example, if i have a maxima - m1, at Gss(:,:,1), i have to find its corresponding maxima at Gss(:,:,2) - (which is just a blurred version of Gss(:,:,1)). How do i do this? I tried to connect it using "distance matrices", if two maxima at different levels are at same position, i link them together. But i find this connection criteria to be very weak and error prone. Are there any other techniques available to solve this problem?