Speaker diarization with early-stop clustering

An apparatus for speaker diarization, the apparatus comprising:

• a segmenting module, for segmenting an audio stream comprising speeches from at least one speaker into at least one speech segment;
• a clustering module for iteratively clustering the at least one speech segment into a plurality of clusters based on cosine similarity and Bayesian Information Criterion (BIC) scores, the number of the plurality of clusters being greater than the number of the at least one speaker;
• a selecting module, for selecting, from the plurality of clusters, at least one cluster of the highest similarity, the number of the selected at least one cluster being equal to the number of the at least one speaker, wherein the selecting is performed by:
• calculating a similarity matrix of the plurality of clusters;
• exhaustively extracting a plurality of similarity sub-matrices from the similarity matrix, each similarity sub-matrix being a MM matrix, M being the number of the at least one speaker;
• calculating an eigen vector of each similarity sub-matrix;
• obtaining an eigenvalue summation of each similarity sub-matrix through summating eigenvalues in an eigen vector of the similarity sub-matrix;
• identifying a similarity sub-matrix having the maximum eigenvalue summation from the plurality of similarity sub-matrices;
• and
• selecting a group of M clusters corresponding to the identified similarity sub-matrix;
• a modeling module, for establishing a speaker classification model based on the selected at least one cluster;
• and
• an aligning module, for aligning, through the speaker classification model, speech frames in the audio stream to the at least one speaker.

An apparatus for speaker diarization, comprising:

• at least one processor;
• and
• a memory storing computer-executable instructions that, when executed, cause the at least one processor to:
• segment an audio stream into at least one speech segment, the audio stream comprising speeches from at least one speaker;
• iteratively cluster the at least one speech segment into a plurality of clusters based on cosine similarity and Bayesian Information Criterion (BIC) scores, the number of the plurality of clusters being greater than the number of the at least one speaker;
• select, from the plurality of clusters, at least one cluster of the highest similarity, the number of the selected at least one cluster being equal to the number of the at least one speaker, by:
• calculating a similarity matrix of the plurality of clusters;
• exhaustively extracting a plurality of similarity sub-matrices from the similarity matrix, each similarity sub-matrix being a MM matrix, M being the number of the at least one speaker;
• calculating an eigen vector of each similarity sub-matrix;
• obtaining an eigenvalue summation of each similarity sub-matrix through summating eigenvalues in an eigen vector of the similarity sub-matrix;
• identifying a similarity sub-matrix having the maximum eigenvalue summation from the plurality of similarity sub-matrices;
• and
• selecting a group of M clusters corresponding to the identified similarity sub-matrix;
• establish a speaker classification model based on the selected at least one cluster;
• and
• align, through the speaker classification model, speech frames in the audio stream to the at least one speaker.