Fusion of a Complementary Feature Set with MFCC for Improved Closed Set Text-Independent Speaker Identification | IEEE Conference Publication | IEEE Xplore