Everybody is good! I want to a 4 d light field data (a four-dimensional matrix) in four directions derivation, using Farid and Simoncelli Derivatives of 3 - tap filter to reduce noise,
P=[0.229879, 0.540242, 0.229879];
D1=[0.425287, 0.000000, 0.425287];
Using 5 - tap filter Derivatives5. M
Used in the image (2 d) matrix by two-dimensional convolution:
Lx=conv2 (p, d1, im, "same");
Ly=conv2 (d1, p, im, "same");
Two-dimensional convolution kernels is by p, d1 of
If I have a four-dimensional matrix, L (u, v, x, y), convolution kernels is supposed to be p, p, p and d1 is a four-dimensional matrix convolution kernels?
View Matlab convn also don't know much about how to use this function on the multidimensional, inquired of god, I how should L in four directions of the partial derivative of Lu, Lv, Lx, Ly

CodePudding user response:

problem of the building Lord too professional, too high!
Don't understand, help,