Generalizing text content summary from speech content

• US 8,868,419 B2
• Filed: 08/22/2011
• Issued: 10/21/2014
• Est. Priority Date: 08/31/2010
• Status: Active Grant

• Current Assignee: Cerence Operating Company (Cerence Inc.)
• Original Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Inventors: Cao, Bao Hua, He, Le, Jin, Xing, Wang, Qing Bo, Zhou, Xin
• Primary Examiner(s): Shah, Paras D
• Assistant Examiner(s): Sharma, Neeraj
• Application Number:
  US13/214,809
• Publication Number:
  US 20120053937A1
• Time in Patent Office: 1,156 Days
• Field of Search: 704/3, 704/7, 704/235, 704270-278, 704/9, 369/25.01, 715/230, 715/232, 715/702, 715/228, 715/254, 715/738, 715/234, 707/913, 707/803, 707/736, 707/754, 707/705, 709/231, 725/46, 725/88, 725/19, 725/61, 370/493, 706/12
• US Class Current: 704/235
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence
• CPC Class Codes: Y10S 707/913 Multimedia

Abstract

A text content summary is created from speech content. A focus more signal is issued by a user while receiving the speech content. The focus more signal is associated with a time window, and the time window is associated with a part of the speech content. It is determined whether to use the part of the speech content associated with the time window to generate a text content summary based on a number of the focus more signals that are associated with the time window. The user may express relative significance to different portions of speech content, so as to generate a personal text content summary.

Claims

1. A system for generating a text content summary from speech content, comprising:

a processor; and

memory connected to the processor, wherein the memory is encoded with instructions and wherein the instructions when executed comprise;

instructions for receiving speech content, the speech content based upon an audio conversation between a first user and a second user;

responsive to reception of at least one focus more signal issued by a the first user, instructions for associating said at least one focus more signal with a first time window, wherein said first time window is associated with a first portion of said speech content, wherein said associating said at least one focus more signal with a first time window further includes determining whether there is currently an active time window;

instructions for determining whether to use said first portion of said speech content associated with said first time window to generate a text content summary based on how many of said at least one focus more signal are associated with said first time window;

instructions for counting via a counting device configured to sum a number of physical user selections of said at least one focus more signal associated with its time window, wherein said counting device also sums a number of physical user selections of said at least one focus less signal associated with its time window, wherein the number of physical user selections corresponds to a number of physical user selections of an input mechanism at an electronic device;

instructions for prioritizing one or more portions of the text content summary based upon, at least in part, the sum of the focus more signal and the sum of the focus less signal; and

instructions for extracting key words from a subset of the one or more portions based on the prioritized portions of the text content summary;

wherein said instructions for determining whether to use said first portion of said speech content comprises instructions for calculating a first focus level value PT, wherein PT=P0+N/T, of said first portion of said speech content associated with said first time window, wherein P0 indicates a default focus level value, T indicates a length of said first time window, and N indicates a number of said at least one focus more signal, and said first focus level value PT of said first portion of said speech content associated with said first time window is ordered so as to use said first portion of said speech content when said first focus level value PT is greater when compared with any other portions of said speech content to generate said text content summary.

2. The system of claim 1, further comprising:

responsive to reception of at least one focus less signal issued by a user, instructions for associating said at least one focus less signal with a second time window, wherein said second time window is associated with a second portion of said speech content; and

instructions for determining whether to use said second portion of said speech content associated with said second time window to generate said text content summary based on how many of said at least one focus less signal are associated with said second time window.

3. The system of claim 2, wherein said instructions for determining whether to use said second portion of said speech content comprises instructions for calculating a second focus level value PT, wherein PT=P0+(N−

M)/T, of said second portion of said speech content associated with said second time window, wherein, P0 indicates a default focus level value, T indicates a length of said second time window, N indicates number of said at least one focus more signal, and M indicates number of said at least one focus less signal, and said second focus level value PT of said second portion of said speech content associated with said second time window is ordered so as to use speech content portions with a relatively larger focus level value PT to generate said text content summary.

4. A method for generating a text content summary from speech content, comprising:

receiving speech content, the speech content based upon an audio conversation between a first user and a second user;

responsive to reception of at least one focus more signal issued by a the first user, associating, using a processor, said at least one focus more signal with a first time window, wherein said first time window is associated with a first portion of said speech content, wherein said associating said at least one focus more signal with a first time window further includes determining whether there is currently an active time window;

determining, using said processor, whether to use said first portion of said speech content associated with said first time window to generate a text content summary based on how many of said at least one focus more signal are associated with said first time window;

counting via a counting device configured to sum a number of physical user selections of said at least one focus more signal associated with its time window, wherein said counting device also sums a number of physical user selections of said at least one focus less signal associated with its time window, wherein the number of physical user selections corresponds to a number of physical user selections of an input mechanism at an electronic device;

prioritizing one or more portions of the text content summary based upon, at least in part, the sum of the focus more signal and the sum of the focus less signal; and

extracting key words from a subset of the one or more portions based on the prioritized portions of the text content summary;

wherein said determining whether to use said first portion of said speech content comprises calculating a first focus level value PT, wherein PT=P0+N/T, of said first portion of said speech content associated with said first time window, wherein P0 indicates a default focus level value, T indicates a length of said first time window, and N indicates a number of said at least one focus more signal, and said first focus level value PT of said first portion of said speech content associated with said first time window is ordered so as to use said first portion of said speech content when said first focus level value PT is greater when compared with any other portions of said speech content to generate said text content summary.

5. The method of claim 4, wherein said associating said at least one focus more signal with a first time window further comprises:

when it is determined there is no currently active time window, generating an active time window, and setting said at least one focus more signal associated with said active time window to 1, wherein said active time window has a start time and an end time, and before arrival of said end time, said active time window is active; and

when it is determined there is a currently active time window, incrementing a number of said at least one focus more signal associated with said active time window by 1.

6. The method of claim 5, further responsive to reception of at least one focus less signal issued by a user:

associating said at least one focus less signal with a second time window, wherein said second time window is associated with a second portion of said speech content; and

determining whether to use said second portion of said speech content associated with said second time window to generate said text content summary further comprises determining a number of said at least one focus less signal associated with said second time window.

7. The method of claim 6, wherein said instructions for determining whether to use said second portion of said speech content comprises instructions for calculating a second focus level value PT, where PT=P0+(N−

M)/T, wherein, P0 indicates a default focus level value, T indicates a length said second time window, N indicates a number of said at least one focus more signal, and M indicates a number of said at least one focus less signal, and said second focus level value PT of said second portion of speech content associated with said second time window are ordered so as to use said speech content portions with a relatively larger focus level value PT to generate a text content summary.

8. A computer program product for generating a text content summary from speech content, the computer program product comprising a non-transitory computer readable storage medium having computer readable program code embodied therewith, the computer readable program code comprising:

computer readable program code configured to receive speech content, the speech content based upon an audio conversation between a first user and a second user;

responsive to reception of at least one focus more signal issued by a the first user, computer readable program code configured to associate said at least one focus more signal with a first time window, wherein said first time window is associated with a first portion of said speech content, wherein said associating said at least one focus more signal with a first time window further includes determining whether there is currently an active time window;

computer readable program code configured to determine whether to use said first portion of said speech content associated with said first time window to generate a text content summary based on how many of said at least one focus more signal are associated with said first time window;

computer readable program code configured to count via a counting device configured to sum a number of physical user selections of said at least one focus more signal associated with its time window, wherein said counting device also sums a number of physical user selections of said at least one focus less signal associated with its time window, wherein the number of physical user selections corresponds to a number of physical user selections of an input mechanism at an electronic device;

computer readable program code configured to prioritize one or more portions of the text content summary based upon, at least in part, the sum of the focus more signal and the sum of the focus less signal; and

computer readable program code configured to extract key words from a subset of the one or more portions based on the prioritized portions of the text content summary;

wherein said computer readable program code configured to determine whether to use said first portion of said speech content comprises computer readable program code configured to calculate a first focus level value PT, wherein PT=P0+N/T, of said first portion of said speech content associated with said first time window, wherein P0 indicates a default focus level value, T indicates a length of said first time window, and N indicates a number of said at least one focus more signal, and said first focus level value PT of said first portion of said speech content associated with said first time window is ordered so as to use said first portion of said speech content when said first focus level value PT is greater when compared with any other portions of said speech content to generate said text content summary.

9. The computer program product of claim 8, wherein said computer readable program code configured to associate said at least one focus more signal with a first time window further comprises:

when it is determined there is no currently active time window, computer readable program code configured to generate an active time window, and set said at least one focus more signal associated with said active time window to 1, wherein said active time window has a start time and an end time, and before arrival of said end time, said active time window is active; and

when it is determined there is a currently active time window, computer readable program code configured to increment a number of said at least one focus more signal associated with said active time window by 1.

10. The computer program product of claim 9, further responsive to reception of at least one focus less signal issued by a user, comprising:

computer readable program code configured to associate said at least one focus less signal with a second time window, wherein said second time window is associated with a second portion of said speech content; and

computer readable program code configured to determine whether to use said second portion of said speech content associated with said second time window to generate said text content summary further comprises computer readable program code configured to determine a number of said at least one focus less signal associated with said second time window.

11. The computer program product of claim 10, wherein said computer readable program code configured to determine whether to use said second portion of said speech content comprises computer readable program code configured to calculate a second focus level value PT, where PT=P0+(N−

M)/T, wherein, P0 indicates a default focus level value, T indicates a length of said second time window, N indicates a number of said at least one focus more signal, and M indicates a number of said at least one focus less signal, and said second focus level value PT of the portions of speech content associated with individual time windows are ordered so as to use speech content portions with relatively larger focus level value PT to generate said text content summary.