mobile radio; principal component analysis; speaker recognition; Euclidian distances; HDM based system; K speakers; LDC CallHome corpus; Mahalanobis based emission model; Mahalanobis distances; data characteristics; hidden distortion-model; several codebooks; spatial layout; speaker diarization; speaker models; speech segment designation; telephone conversations; Covariance matrices; Density estimation robust algorithm; Hidden Markov models; Speech; Standards; Training; Vectors; Hidden-distortion model (HDM); K-means; Mahalanobis distance; self-organizing maps (SOM); speaker diarization;
机译:电话会议基于迭代的说话人区分系统的初始化
机译:结合高斯化/非高斯化功能以改善电话对话中的说话人差异化
机译:基于通用维特比的时间序列分割和聚类模型,用于说话人区分
机译:基于马哈拉诺比斯的发射模型,用于电话对话中的说话人区分
机译:基于对话的说话人辨别的模型形成和分类技术。
机译:使用预训练的视听同步模型进行多模态扬声器二分法
机译:多功能组合改善电话对话的扬声器化
机译:麻省理工学院林肯实验室RT-04F Diarization systems:广播音频和电话对话的应用