Identical conversation detection method and apparatus

• US 9,621,713 B1
• Filed: 04/01/2014
• Issued: 04/11/2017
• Est. Priority Date: 04/01/2014
• Status: Active Grant

• Current Assignee: Securus Technologies Incorporated (Securus Special Investments LLC)
• Original Assignee: Securus Technologies Incorporated (Securus Special Investments LLC)
• Inventors: Gainsboro, Jay, Sinclair, Ken, Weinstein, Lee
• Primary Examiner(s): Bezuayehu, Solomon
• Application Number:
  US14/242,026
• Time in Patent Office: 1,106 Days
• Field of Search: 37921002-21003, 379 8801- 8819
• CPC Class Codes: G10L 17/00 Speaker identification or verification techniques
• CPC Class Codes: G10L 25/51 for comparison or discrimination
• CPC Class Codes: H04L 67/535 Tracking the activity of the user network monitoring arrangements H04L43/00; recording of computer activity G06F11/34
• CPC Class Codes: H04M 3/2281 Call monitoring, e.g. for law enforcement purposes; Call tracing; Detection or prevention of malicious calls
• CPC Class Codes: H04M 3/42221 Conversation recording systems at the subscriber's set H04M1/656

Abstract

An automated system is disclosed for detecting situations in which identical segments of conversation appear within two different recordings. The system automatically detects where within each of two audio recordings an identical conversation segment begins and ends, thus enabling detection of conversations where multiple prison inmates participated in a conference call.

Claims

1. A method of detecting unauthorized communications between two incarcerated persons, comprising:

monitoring a plurality of communications, each of the monitored communications originated from a communication device;

generating a time-segmented call signal from each of the plurality of monitored communications;

storing in a database the dialed number for each monitored communication, the start time of each monitored communication, and the duration or end time of each monitored communication;

identifying a first communication and a second communication in the database of monitored communications, where the identified first and second communications at least partially overlap in time;

generating a first time-segmented call signature for a portion of the first communication that overlaps with the second communication and further generating a second time-segmented call signature for a portion of the second communication that overlaps with the first communication;

analyzing the first time-segmented call signature and the second time-segmented call signature to identify identical segment signature data in an overlapping portion of the first communication and the second communication, wherein the first time-segmented call signature and the second time-segmented call signature are analyzed using a digital signal processor; and

alerting a prison official when identical segment signature data is identified in the overlapping communications.

2. The method of claim 1, further comprising:

detecting whether the first communication and the second communication are placed to the same dialed number.

3. The method of claim 1, wherein the first time-segmented call signature and the second time-segmented call signature are generated based on identified spectral peaks in the monitored communications.

4. The method of claim 1, wherein the time boundaries of the first time-segmented signature and the second time-segmented signature are derived based on detected features of the monitored communications.

5. The method of claim 3, wherein the call signature data is invariant to an absolute time.

6. The method of claim 4, further comprising:

indexing the time-segmented signals, wherein each time-segmented signal is indexed based on one or more hashes generated based on the time-segmented signal; and

storing the time-segmented signals in the database based on the generated hashes.

7. The method of claim 6, wherein the identical segment signature data is identified based on the hashes uses to index the time-segmented signals.

8. The method of claim 1, wherein the first time-segmented call signature and the second time-segmented call signature begin at different absolute times.

9. The method of claim 1, wherein the first time-segmented call signature and the second time-segmented call signature are insensitive to the amplitude of the monitored communications.

10. The method of claim 1, wherein the first communication and the second communication are recorded conversations.

11. A system for detecting unauthorized communications between two incarcerated persons, the system comprising:

a processor; and

a memory coupled to the processor, the memory storing computer-readable instructions that, upon execution by the processor, cause the system to;

monitor a plurality of communications, each of the monitored communications originated from a communication device;

generate a time-segmented call signal from each of the plurality of monitored communications;

store in a database the dialed number for each monitored communication, the start time of each monitored communication, and the duration or end time of each monitored communication;

identify a first communication and a second communication in the database of monitored communications, where the identified first and second communications at least partially overlap in time;

generate a first time-segmented call signature for a portion of the first communication that overlaps with the second communication and further generating a second time-segmented call signature for a portion of the second communication that overlaps with the first communication;

analyze the first time-segmented call signature and the second time-segmented call signature to identify identical segment signature data in an overlapping portion of the first communication and the second communication, wherein the first time-segmented call signature and the second time-segmented call signature are analyzed using a digital signal processor; and

alert a prison official identical segment signature data is identified in the overlapping communications.

12. The system of claim 11, wherein the memory storing additional computer-readable instructions that, upon execution by the processor, cause the system to:

detect whether the first communication and the second communication are placed to the same dialed number.

13. The system of claim 11, wherein the first time-segmented call signature and the second time-segmented call signature are generated based on identified spectral peaks in the monitored communications.

14. The system of claim 11, wherein the time boundaries of the first time-segmented signature and the second time-segmented signature are derived based on detected features of the monitored communications.

15. The system of claim 13, wherein the call signature data is invariant to an absolute time.

16. The system of claim 14, wherein the memory storing additional computer-readable instructions that, upon execution by the processor, cause the system to:

index the time-segmented signals, wherein each time-segmented signal is indexed based on one or more hashes generated based on the time-segmented signal; and

store the time-segmented signals in the database based on the generated hashes.

17. The system of claim 16, wherein the identical segment signature data is identified based on the hashes uses to index the time-segmented signals.

18. The system of claim 11, wherein the first time-segmented call signature and the second time-segmented call signature begin at different absolute times.

19. The system of claim 11, wherein the first time-segmented call signature and the second time-segmented call signature are insensitive to the amplitude of the monitored communications.

20. The system of claim 11, wherein the first communication and the second communication are recorded conversations.