Unsupervised speaker segmentation of multi-speaker speech data

A system for segmenting speech data into speaker segments by speaker, the system comprising:

• a processor;
• a first module configure to control the processor to scan input speech data with a windowed generalized likelihood ratio (GLR) function to obtain speech segments, wherein the input speech data includes a plurality of speakers;
• a second module configured to control the processor to cluster the speech segments into one or more clusters, wherein each cluster is associated with a single speaker;
• a third module configured to control the processor, if more clusters exist than speakers, to:
• check overlap between segments in each cluster;
• pool clusters that have overlap between at least one segment in each pooled cluster;
• and
• resegment and remodel the pooled clusters;
• a fourth module configured to control the processor to create models for each cluster;
• and
• a fifth module configured to control the processor to rescan the input speech data with the models to resegment the speech data and obtain speech segments for each speaker included in the speech data.

A system for segmenting speech data by speaker, the system comprising:

• a processor;
• a first module configured to control the processor to obtain initial estimated segments of the speech data, wherein the estimated segments are unlabeled;
• a second module configured to control the processor to cluster the initial estimated segments until the initial estimated segments are grouped into a final group;
• a third module configured to control the processor to select one or more clusters from the final group, wherein one or more clusters corresponds to groups of estimated segments;
• a fourth module configured to control the processor to iteratively model and resegment each group of estimated segments until changes in segment boundaries for the estimated segments in each group of estimated segments from a particular iteration to a next iteration are below a threshold;
• a fifth module configured to control the processor, if at least one segment in a first group of estimated segments overlaps with at least one segment in a second group of estimated segments, to pool the first group and the second group of estimated segments;
• a sixth module configured to control the processor to iteratively model and resegment the pooled first and second group and any remaining groups of estimated segments to generate new groups of estimated segments until the overlap between the new groups of estimated segments is below a particular level;
• and
• a seventh module configured to control the processor to associate each group of estimated segments and each new group of estimated segments, if any, with a different speaker.