System and method of video capture and search optimization for creating an acoustic voiceprint

1. A method for creating an acoustic voiceprint for a specific speaker, the method comprising:

providing a diarization system for creating and applying acoustic voiceprints to audio data, wherein the diarization system includes an audio database server, a voiceprint database server, a processor, and an audio source, wherein the processor is operably connected to the audio database server and the voiceprint database server;

receiving, by the processor, a plurality of audio files from the audio database server; receiving, by the processor, agent identification metadata associated with each of the

plurality of audio files from the audio database server;

selecting, by the processor, a subset of audio files from the plurality of audio files based on the agent identification metadata, wherein the subset of audio files belong to a specific agent;

diarizing, by the processor, each of the subset of audio files by:

separating the audio file into a plurality of speaker segments of speech separated by non-speech,

clustering similar speaker segments of the plurality of segments, wherein the similar speaker segments have a high likelihood of containing speech from a single speaker, and

classifying each of the clustered speaker segments as belonging to the specific agent or as belonging to an other speaker;

building, by the processor, an acoustic voiceprint for the specific agent using the clustered speaker segments classified as belonging to the specific agent;

saving the acoustic voiceprint to the voiceprint database server;

applying, by the processor, the acoustic voiceprint to a new audio file, received from the audio source, to determine which segments match the acoustic voiceprint; and

labeling, by the processor, the segments of the new audio file that match the acoustic voiceprint as belonging to the specific agent.

2. The method of claim 1 , wherein clustering the similar speaker segments of the plurality of segments includes applying at least one metric to the segments with the processor to label the segments of speech as belonging to the specific agent or as belonging to the other speaker.

3. The method of claim 2 , wherein the at least one metric is that of cluster size wherein the larger the cluster the more likely the segment belongs to the specific agent.

4. The method of claim 1 , wherein the selected subset of audio files are also the top 50% or less of audio files belonging to the specific agent based on an acoustical difference between the specific agent and the other speaker.

5. The method of claim 1 , wherein the selected subset of audio files are also the top 20% or less of audio files belonging to the specific agent based on an acoustical difference between the specific agent and the other speaker.

6. The method of claim 1 , wherein the selected subset of audio files are further selected by sending the audio files belonging to the specific speaker to a transcription server wherein the transcription server transcribes the audio files and automatedly scores each transcribed audio file for a confidence of transcription, only those audio files that maximize the confidence of transcription and maximize an acoustical difference in voice frequencies between the known speaker and the at least one other speaker in the same audio file are selected.

7. The method of claim 1 , the method further comprising:

performing a blind diarization on the new audio file to separate the new audio file into at least a first speaker audio file and a second speaker audio file;

wherein the acoustic voiceprint is applied to the first speaker audio file and the second speaker audio file to determine if one of the speaker audio files is the specific agent.

8. The method of claim 7 , the method further comprising:

receiving speaker metadata for the new audio file; and

selecting the acoustic voiceprint to apply to the new audio file from a plurality of acoustic voiceprints based upon the received speaker metadata.

9. The method of claim 8 , wherein the blind diarization is based in part upon the acoustic voiceprint.

10. The method of claim 1 , wherein the new audio file is real time

audio data.

11. A system for creating an acoustic voiceprint for a specific speaker, the system comprising:

a diarization system for creating and applying acoustic voiceprints to audio data, wherein the diarization system includes:

an audio database including a plurality of audio files and agent identification metadata associated with each of the plurality of audio files;

a voiceprint database server for storing acoustic voiceprints;

an audio source; and

a processor, wherein the processor is operably connected to the audio database server and the voiceprint database server, wherein the processor:

receives a plurality of audio files from the audio database server,

receives agent identification metadata associated with each of the plurality of audio files from the audio database server,

selects a subset of audio files from the plurality of audio files based on the agent identification metadata, wherein the subset of audio files belong to a specific agent,

diarizes each of the subset of audio files by:

separating the audio file into a plurality of speaker segments of speech separated by non-speech,

clustering similar speaker segments of the plurality of segments, wherein the similar speaker segments have a high likelihood of containing speech from a single speaker, and

classifying each of the clustered speaker segments as belonging to the specific agent or as belonging to an other speaker,

builds an acoustic voiceprint for the specific agent using the clustered speaker segments classified as belonging to the specific agent,

save the acoustic voiceprint to the voiceprint database server,

applies the acoustic voiceprint to a new audio file, received from the audio source, to determine which segments match the acoustic voiceprint, and

labels the segments of the new audio file that match the acoustic voiceprint as belonging to the specific agent.

12. The system of claim 11 , wherein clustering the similar speaker segments of the plurality of segments includes applying at least one metric to the segments with the processor to label the segments of speech as belonging to the specific agent or as belonging to the other speaker.

13. The system of claim 12 , wherein the at least one metric is that of cluster size wherein the larger the cluster the more likely the segment belongs to the specific agent.

14. The system of claim 11 , wherein the selected subset of audio files are also the top 50% or less of audio files belonging to the specific agent based on an acoustical difference between the specific agent and the other speaker.

15. The system of claim 11 , wherein the selected subset of audio files are also the top 20% or less of audio files belonging to the specific agent based on an acoustical difference between the specific agent and the other speaker.

16. The system of claim 11 , wherein the selected subset of audio files are further selected by sending the audio files belonging to the specific speaker to a transcription server wherein the transcription server transcribes the audio files and automatedly scores each transcribed audio file for a confidence of transcription, only those audio files that maximize the confidence of transcription and maximize an acoustical difference in voice frequencies between the known speaker and the at least one other speaker in the same audio file are selected.

17. The system of claim 11 , wherein the processor further:

performs a blind diarization on the new audio file to separate the new audio file into at least a first speaker audio file and a second speaker audio file;

wherein the acoustic voiceprint is applied to the first speaker audio file and the second speaker audio file to determine if one of the speaker audio files is the specific agent.

18. The system of claim 17 , wherein the processor further:

receives speaker metadata for the new audio file; and

selects the acoustic voiceprint to apply to the new audio file from a plurality of acoustic voiceprints based upon the received speaker metadata.

19. The system of claim 18 , wherein the blind diarization is based in part upon the acoustic voiceprint.

20. The system of claim 11 , wherein the new audio file is real time audio data.