Caller identification in a secure environment using voice biometrics
View Patent ↗A method of providing a transcription of an electronic communication includes determining a language spoken in the electronic communication, differentiating among participants in the electronic communication, identifying a hierarchy of topics discussed in the electronic communication, and providing a display of the hierarchy of topics.
1. A method of providing a transcription of an electronic communication, comprising:
using a language identification engine to determine a language spoken in the electronic communication;
using a participation identification engine to differentiate among participants in the electronic communication;
using a topic identification engine to identify a hierarchy of topics discussed in the electronic communication and transcribe the electronic communication using the determined language by converting the electronic communication to text and applying participant segmentation analysis and time alignment methods that delineate when one or another individual is speaking;
implementing the participation identification engine, language identification engine, and topic identification engine using a machine learning model trained using known participant biometric voice prints, predetermined language constructs, and a known set of topics; and
providing a display of the transcription, the participants, the determined language, and the hierarchy of topics.
2. The method of claim 1 , wherein the electronic communication is a live or recorded electronic communication between one or more participants within a corrections facility and one or more participants outside of the corrections facility.
3. The method of claim 2 , wherein the live or recorded electronic communication is at least one of a telephone call, video call, email, text message, streaming audio and video, or audio loaded from outside sources.
4. The method of claim 1 , wherein the language is selected from a library of languages.
5. The method of claim 4 , wherein the language is selected based on a best match of one or more of intonations, phonetic pronunciations, and vocabulary.
6. The method of claim 1 , comprising inserting a tone in a particular participant's portion of the electronic communication to differentiate among the participants.
7. The method of claim 1 , comprising identifying the topics using rule based criteria.
8. The method of claim 1 , comprising identifying the topics using user defined criteria.
9. The method of claim 1 , comprising identifying the topics by indexing words of the electronic communication in a lattice matrix database with a best match and alternatives for each word.
10. The method of claim 1 , wherein the display of the hierarchy of topics illustrates most pronounced topics in a larger font.
11. The method of claim 1 , comprising providing a searching facility to search the electronic communication for one or more of:
a topic;
a particular word, phrase, sentence, or paragraph;
a particular participant's voice
a particular language;
a particular destination number;
a particular participant's gender.
12. The method of claim 1 , comprising providing alerts based on one or more criteria comprising:
a topic;
a particular word, phrase, sentence, or paragraph;
a particular participant's voice
a particular language;
a particular destination number;
a particular participant's gender.
13. The method of claim 1 , comprising operating the participation identification engine to analyze electronic conversations originating from particular participants to differentiate among the participants, the electronic conversations satisfying attributes including one or more of devices used for the electronic conversations, time of day, duration, an amount of time during which the participants speak, number of detected participants, signal to noise ratios, and saturation levels.
14. The method of claim 1 , wherein the language identification engine comprises a natural language processor configured to perform automatic language identification of languages spoken in the electronic communication and translate the identified languages to a common language using a language library.
15. The system of claim 14 , comprising selecting a limited number of languages for use by the language library to lower computing resources required for language identification and to provide precise translation.
16. A system for providing a transcription of an electronic communication, comprising:
a processor;
a computer readable medium storing computer readable program code, that when executed by the processor, causes the processor to implement:
a participation identification engine for differentiating among participants in the electronic communication;
a language identification engine for determining a language spoken in the electronic communication; and
a topic identification engine for identifying a hierarchy of topics discussed in the electronic communication and for transcribing the electronic communication using the determined language by converting the electronic communication to text and applying participant segmentation analysis and time alignment methods that delineate when one or another individual is speaking,
wherein the participation identification engine, language identification engine, and topic identification engine are implemented using a machine learning model trained using known participant biometric voice prints, predetermined language constructs, and a known set of topics; and
a user interface for providing a display of the transcription, the participants, the determined language, and the hierarchy of topics.
17. The system of claim 16 , wherein the electronic communication is a live or recorded electronic communication between one or more participants within a corrections facility and one or more participants outside of the corrections facility.
18. The system of claim 17 , wherein the live or recorded electronic communication is at least one of a telephone call, video call, email, text message, streaming audio and video, or audio loaded from outside sources.
19. The system of claim 16 , wherein the language is selected from a library of languages.
20. The system of claim 19 , wherein the language is selected based on a best match of one or more of intonations, phonetic pronunciations, and vocabulary.
21. The system of claim 16 , wherein the participation identification engine operates to insert a tone in one participant's portion of the electronic communication to differentiate among the participants.
22. The system of claim 16 , wherein the topic identification engine operates to identify the topics using a rule based criteria.
23. The system of claim 16 , wherein the topic identification engine operates to identify the topics using user defined criteria.
24. The system of claim 16 , wherein the topic identification engine operates to identify the topics by indexing words of the electronic communication in a lattice matrix database with a best match and alternatives for each word.
25. The system of claim 16 , wherein the topic identification engine causes the user interface to display the most pronounced topics of the hierarchy of topics in a larger font.
26. The system of claim 16 , wherein the topic identification engine operates to provide a searching facility to search the electronic communication for one or more of:
a topic;
a particular word, phrase, sentence, or paragraph;
a particular participant's voice
a particular language;
a particular destination number;
a particular participant's gender.
27. The system of claim 16 , wherein the topic identification engine operates to provide alerts based on one or more criteria comprising:
a topic;
a particular word, phrase, sentence, or paragraph;
a particular participant's voice
a particular language;
a particular destination number;
a particular participant's gender.
28. The system of claim 16 , wherein the participation identification engine operates to analyze electronic conversations originating from particular participants to differentiate among the participants, the electronic conversations satisfying attributes including one or more of devices used for the electronic conversations, time of day, duration, an amount of time during which the participants speak, number of detected participants, signal to noise ratios, and saturation levels.
29. The system of claim 16 , wherein the language identification engine comprises a natural language processor configured to perform automatic language identification of languages spoken in the electronic communication and translate the identified languages to a common language using a language library.
30. The system of claim 29 , wherein the language library comprises a limited number of languages selected to lower computing resources required for language identification and to provide precise translation.