Source-normalised-and-weighted LDA for robust speaker recognition using i-vectors.

Mitchell McLaren, David A. van Leeuwen

Browse the full ICASSP paper archive.