Speaker recognition evaluation: selective approaches and fusion

From LRDE

Abstract

FIXME