Approximation Capabilities of Neural Networks using Morphological Perceptrons and Generalizations