Detects gender of the speaker based on various features like MFCC, pitch, short-time energy, energy entropy, zero-crossing rate and spectral centroid. Also contains an android application to detect gender based on the pitch of a person's voice. Uses different kinds of filters (mean, median, mode). The former uses a machine learning approach while the latter is a more direct and naive approach.
Find the detailed report here.