Fig. 5From: The use of long-term features for GMM- and i-vector-based speaker diarization systemsProposed i-vector-based clustering architecture. The proposed i-vector-based cosine-distance and PLDA clustering are based on short- and long-term speech featuresBack to article page