Use mirtk::ScalarGaussian of Numerics module

[schuhschuh]

Remove mirtkGaussian.h and mirtkGaussian.cc files and use mirtk::ScalarGaussian class from Numerics module instead which implements the exact same thing. Possibly add missing default argument .0 for y_0 in argument list of constructor of mirtk::ScalarGaussian such that you can write

mirtk::ScalarGaussian gaussian(sigma, mu);

or add another constructor overload with two double arguments where the order is "mu, sigma".