Analytical Solution of the Blind Source Separation Problem Using Derivatives