Auditory-inspired Morphological Processing of Spectrograms for Speech Processing