Detecting an answering machine using speech recognition

• US 20080015846A1
• Filed: 07/12/2006
• Published: 01/17/2008
• Est. Priority Date: 07/12/2006
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Ju, Yun-Cheng, Yu, Dong, Acero, Alejandro, Wang, Ye-Yi, Fisher, Craig M.
• Granted Patent: US 8,065,146 B2
• Time in Patent Office: Days
• US Class Current: 704/201
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence
• CPC Class Codes: G10L 25/78 Detection of presence or absence of voice signals switching of direction of transmission by voice frequency in two-way loud-speaking telephone systems H04M9/10
• CPC Class Codes: H04M 2203/2027 Live party detection
• CPC Class Codes: H04M 3/5158 in combination with automated outdialling systems devices for signalling identity of wanted subscriber H04M1/26

Abstract

An answering machine detection module is used to determine whether a call recipient is an actual person or an answering machine. The answering machine detection module includes a speech recognizer and a call analysis module. The speech recognizer receives an audible response of the call recipient to a call. The speech recognizer processes the audible response and provides an output indicative of recognized speech. The call analysis module processes the output of the speech recognizer to generate an output indicative of whether the call recipient is a person or an answering machine.

Claims

1. A computer-implemented method of ascertaining whether a call recipient is an actual person or an answering machine, the method comprising:

receiving an audible response from a call recipient and processing the audible response with a speech recognizer to provide an output indicative of recognized speech; and

processing the output indicative of recognized speech with a classifier to provide an output indicative of whether the call recipient is an actual person or an answering machine, said processing being based on a statistical analysis of the output indicative of recognized speech.

2. The computer-implemented method of claim 1 wherein processing includes processing the audible response to ascertain non-word features associated with the audible response and using the non-word features as a basis to provide an output indicative of whether the call recipient is an actual person or an answering machine.

3. The computer-implemented method of claim 2 wherein processing includes ascertaining a duration of the audible response and using the duration of the audible response as a basis to provide an output indicative of whether the call recipient is an actual person or an answering machine.

4. The computer-implemented method of claim 2 wherein processing includes ascertaining whether silence is initially present in the audible response and using the presence of silence as a basis to provide an output indicative of whether the call recipient is an actual person or an answering machine.

5. The computer-implemented method of claim 4 wherein processing includes ascertaining whether silence of a period relative to a selected length is initially present in the audible response and using the presence of silence as a basis to provide an output indicative of whether the call recipient is an actual person or an answering machine.

6. The computer-implemented method of claim 2 wherein processing includes using an indication as to whether the speech recognizer was able to recognize any speech in the audible response as a basis to provide an output indicative of whether the call recipient is an actual person or an answering machine.

7. The computer-implemented method of claim 3 wherein processing includes using statistical analysis when the duration of the utterance is within a selected range.

8. The computer-implemented method of claim 1 wherein processing includes using statistical analysis when a duration of the utterance is within a selected range.

9. A computer readable medium having instructions which when implemented by a computer ascertain whether a call recipient is an actual person or an answering machine, the instructions comprising:

receiving an audible response from a call recipient and processing the audible response with a speech recognizer to provide an output indicative of recognized speech; and

processing the output indicative of recognized speech to provide an output indicative of whether the call recipient is an actual person or an answering machine, said processing being based on analysis of word phrases in the output indicative of recognized speech.

10. The computer readable medium of claim 9 wherein processing comprises processing the word phrases with a statistical classifier and using an output from the statistical classifier as a basis to provide an output indicative of whether the call recipient is an actual person or an answering machine.

11. The computer readable medium of claim 9 wherein processing includes processing the audible response to ascertain non-word features associated with the audible response and using the non-features as a basis to provide an output indicative of whether the call recipient is an actual person or an answering machine.

12. The computer readable medium of claim 9 wherein processing includes ascertaining a duration of the audible response and using the duration of the audible response as a basis to provide an output indicative of whether the call recipient is an actual person or an answering machine.

13. The computer readable medium of claim 9 wherein processing includes ascertaining whether silence is initially present in the audible response and using the presence of silence as a basis to provide an output indicative of whether the call recipient is an actual person or an answering machine.

14. The computer readable medium of claim 13 wherein processing includes ascertaining whether silence of a period relative to a selected length is initially present in the audible response and using the presence of silence as a basis to provide an output indicative of whether the call recipient is an actual person or an answering machine.

15. The computer readable medium of claim 9 wherein processing includes using an indication as to whether the speech recognizer was able to recognize any speech in the audible response as a basis to provide an output indicative of whether the call recipient is an actual person or an answering machine.

16. The computer readable medium of claim 10 wherein processing includes using statistical analysis when the duration of the utterance is within a selected range.

17. The computer readable medium of claim 10 and further comprising operating the speech recognizer to detect barge-in events by the call recipient and replaying a message to the call recipient if a barge-in event is detected until the message is played in its entirety.

18. The computer readable medium of claim 17 wherein replaying the message to the call recipient comprises replaying the message to the call recipient if the call recipient is the answering machine.

19. A computer-implemented method of leaving a message on an answering machine, the method comprising:

operating a speech recognizer to detect barge-in events by the answering machine; and

restarting a message to the answering machine if a barge-in event is detected until the message is played in its entirety.

20. The computer implemented method of claim 19 and further comprising, prior to the step of replaying the message, detecting whether a call recipient is an actual person or the answering machine wherein detecting comprises:

receiving an audible response from the call recipient and processing the audible response with the speech recognizer to provide an output indicative of recognized speech; and

processing the output indicative of recognized speech to provide an output indicative of whether the call recipient is an actual person or the answering machine, said processing being based on analysis of word phrases in the output indicative of recognized speech.