Method and apparatus for extracting voiced telephone numbers and email addresses from voice mail messages

• US 6,785,367 B2
• Filed: 03/18/2002
• Issued: 08/31/2004
• Est. Priority Date: 03/20/2001
• Status: Expired due to Term

• Current Assignee: Mitel Networks Corporation
• Original Assignee: Mitel Knowledge Corporation (Mitel Networks Corporation)
• Inventors: Horvath, Steve, Kasvand, Tonis
• Primary Examiner(s): WEAVER, SCOTT LOUIS
• Application Number:
  US10/101,083
• Publication Number:
  US 20020196910A1
• Time in Patent Office: 897 Days
• Field of Search: 379/67.1, 379/88.01, 379/88.02, 379/88.03, 379/88.04, 379/88.08, 379/88.19, 379/88.22, 704/270, 704/275, 704/276
• US Class Current: 379/88.04
• CPC Class Codes: H04M 2201/40 using speech recognition speech recognition per se G10L15/00
• CPC Class Codes: H04M 2203/301 Management of recordings
• CPC Class Codes: H04M 3/53333 Message receiving aspects
• CPC Class Codes: H04M 3/5335 Message type or catagory, e.g. priority, indication

Abstract

The present invention relates to a system for automatically identifying enunciation of telephone numbers and email addresses in voice mail messages for presentation to a user via either voice or text, without the user being required to replay the message. Furthermore, according to the invention the extracted telephone number or email addresses may be forwarded to an application for further processing (e.g. to dial the extracted number, add the email address and/or telephone number to a contacts database, etc.) The user may specify text strings to be searched (other than telephone numbers and email addresses). Thus, a user can specify certain words such as “extension”, “email”, etc., in order to qualify the extracted number or email address information, or words such as “urgently” or “immediately” to identify important voice mail messages.

Claims

1. A system for extracting information from a voicemail message, comprising:

a knowledge database for storing sound and noise algorithms representing symbols, numbers, letters, and words arranged according to user programmed grammar representations representing phrases to be identified in said voicemail message; and

a voiced number extraction application for dividing said voicemail message into detected speech patterns, comparing said speech patterns to said sound and noise algorithms using said grammar representations and extracting and presenting matching ones of said symbols, numbers, letters, and words corresponding to said phrases to be identified in said voicemail message;

wherein each of said phrases comprises at least one hot symbol matching one of said speech patterns in order to identify one of said phrases.

2. The system of claim 1, further including a user interface for launching said application, accepting said phrases identified in said voicemail message and initiating predetermined contextual actions based thereon, and for terminating said application.

3. The system of claim 2, wherein said user interface selectively displays multiple ones of said contextual actions for associated ones of said phrases.

4. The system of claim 1, further including a user interface for entering said grammar representations representing phrases to be identified in said voicemail message.

5. The system of claim 1, wherein each of said phrases further comprises a wildcard representing any symbol which is not required to match one of said speech patterns in order to identify said one of said phrases.

6. The system of claim 5, wherein said wildcard is a number wildcard representing any number which is not required to match one of said speech patterns in order to identify said one of said phrases.

7. A method of extracting information from a voicemail message, comprising:

storing sound and noise algorithms representing symbols, numbers, letters, and words arranged according to user programmed grammar representations representing phrases to be identified in said voicemail message, wherein each of said phrases comprises at least one hot symbol;

dividing said voicemail message into detected speech patterns;

comparing said speech patterns to said sound and noise algorithms using said grammar representations and extracting and presenting matching ones of said symbols, numbers, letters, and words corresponding to said phrases to be identified in said voicemail message, wherein said at least one hot symbol matches one of said speech patterns in order to identify one of said phrases.

8. The method of claim 7, further comprising initiating predetermined contextual actions based on said phrases identified in said voicemail message.

9. The method of claim 8, further comprising selectively displaying multiple ones of said contextual actions for associated ones of said phrases.

10. The method of claim 9, wherein said comparing further comprises:

i) storing a predetermined number of said detected speech patterns in respective FIFO memories, wherein said predetermined number is equal to the number of symbols, numbers, letters, and words in the longest of said phrases to be identified;

ii) initializing a plurality of tracking pointers to respective ones of said symbols, numbers, letters, and words in respective ones of said phrases to be identified;

iii) comparing successive ones of said symbols, numbers, letters, and words identified by said tracking pointers to said speech patterns stored in said FIFO memories;

iv) until each of said symbols, numbers, letters, and words has been compared, increment said tracking pointers and repeat step iii);

v) until each detected speech pattern has been compared, shift said detected speech patterns through said FIFO memories and repeat steps ii) through iv); and

vi) in the event of a match display said match.