Principle curvature to gaussian curvatures, or gaussian curvature to principle curvature? which is simple?

Which is simple if I want to calculate surface curvature.

By reading this:

(1)

The principal curvature values and the principle directions of the curvatures are calculated by taking the eigenvalue decomposition of the Hessian Matrix which is defined as:

Using principal curvatures, Mean (H) and Gaussian (K) Curvatures are calculated as:
H = (K 1 +K 2 )/ 2 , K =k1 ⋅k 2

(2) method

(2) method is:

from http://users.vcnet.com/simonp/curvature.pdf

Which is easy to calculate in matlab?