Fusing MFCC and LPC Features Using 1D Triplet CNN for Speaker Recognition in Severely Degraded Audio Signals | IEEE Journals & Magazine | IEEE Xplore