Fig. 8From: Latent class model with application to speaker diarizationDER varies with k of soft prior (20)Back to article page