Matlab Data:

Variables stored in imgproc.mat

img1 (1024 x 1024): Output from the Mejiering ridge detector (maximg in Aurora’s code)

img2 (1024 x 1024): Output from Automated Neurite….. paper’s “ridge detection” (using img1 as the input)

-These are the minimum eigenvalues across scales (e.g. sigma = 1, 2, 3, 4, 5, 6, 7)

img3 (1024 x 1024): Output from Automated Neurite….. paper’s “ridge detection” (using img2 as the input)

-Again, minimum eigenvalues across scales

img2_eigvec (1024 x 1024 x 2):

img3_eigvec (1024 x 1024 x 2):

If the eigenvector associated with each point is [nx_i,j ny_i,j]^T, where i = row index, j = column index,

then say for the pixel at i = 1, j = 1 for img2,

nx_1,1 = img2_eigvec(1,1,1);

ny_1,1 = img2_eigvec(1,1,2);

-img2_eigvec(:,:,1) is the first entry of the eigenvector (nx_i,j for all i, j)

-img2_eigvec(:,:,2) is the second entry of the eigenvector (ny_i,j for all i, j)

To access all nx, use squeeze(img2_eigvec(:,:,1))

To access all ny, use squeeze(img2_eigvec(:,:,2))