Apparatus and method for automatically generating punctuation marks in continuous speech recognition

• US 6,718,303 B2
• Filed: 05/13/1999
• Issued: 04/06/2004
• Est. Priority Date: 05/13/1998
• Status: Expired due to Term

• Current Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Original Assignee: International Business Machines Corporation
• Inventors: Tang, Donald T., Zhu, Xiao Jin, Shen, Li Oin
• Primary Examiner(s): Dorvil, Richemond
• Assistant Examiner(s): ABEBE, DANIEL DEMELASH
• Application Number:
  US09/311,503
• Publication Number:
  US 20020069055A1
• Time in Patent Office: 1,790 Days
• Field of Search: 704/235, 704/251, 704/256, 704/246, 704/215, 704/233, 704/257
• US Class Current: 704/235
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence

Abstract

An apparatus for automatically generating punctuation marks in a continuous speech recognition system, comprises means (1,2,3,5) for recognizing user speech and converting the user speech into words, characterized in that means (1,2,3,5) for recognizing user speech is further used to recognize pseudo noises in the user speech; and the apparatus characterized by further comprising: means (9) for marking pseudo noises in output results of means (1,2,3,5) for recognizing user speech; means (10,14,13) for generating punctuation marks by finding most likely pseudo punctuation marks at locations of pseudo noises marked by the means (9) for marking pseudo noises based on a language model containing pseudo punctuation marks.

Claims

1. An apparatus for automatically generating punctuation marks in a continuous speech recognition system, comprising means for recognizing user speech and converting said user speech into words,

said means for recognizing user speech further recognizing pseudo noises in said user speech;

and

said apparatus further comprising;

means for marking pseudo noises in output results of said means for recognizing user speech; and

means for generating punctuation marks by finding most likely pseudo punctuation marks at locations of pseudo noises marked by means for marking pseudo noises based on a language model containing pseudo punctuation marks, said means for generating punctuation marks comprising;

means for calculating probability for each pseudo punctuation mark in a language model containing pseudo punctuation marks that said pseudo punctuation mark would occur in said output results of said means for recognizing user speech;

means for calculating probabilities that particular pseudo noises will occur at locations of particular pseudo punctuation marks; and

means for finding most likely pseudo punctuation marks at locations of said pseudo noises marked by said means for marking pseudo noises based on said probabilities calculated, and generating punctuation marks corresponding to said most likely pseudo punctuation marks.

2. An apparatus for automatically generating punctuation marks in a continuous speech recognition system, comprising means for recognizing user speech and converting said user speech into words, further comprising:

means for generating location indicating signals in response to operations of a user during dictation, said location indicating signals indicating locations in output results of said means for recognizing user speech;

means for calculating probability for each pseudo punctuation mark in a language model containing pseudo punctuation marks that said pseudo punctuation mark would occur in said output results of said means for recognizing; and

means for finding most likely pseudo punctuation marks at locations indicated by said location indicating signal based on said probabilities calculated, and generating punctuation marks corresponding to said most likely pseudo punctuation marks.

3. A method for automatically generating punctuation marks in a continuous speech recognition system, comprising a step of recognizing user speech and converting said user speech into words,

pseudo noises in said user speech are also recognized in said step of recognizing user speech;

and

said method further comprising steps of;

marking pseudo noises in output results of said step of recognizing user speech; and

generating punctuation marks by finding most likely pseudo punctuation marks at locations of said pseudo noises marked in said step of marking pseudo noises, based on a language model containing pseudo punctuation marks, said step of generating punctuation marks comprising the steps of;

calculating probability for each pseudo punctuation mark in a language model containing pseudo punctuation marks that said pseudo punctuation mark would occur in said output results of said step of recognizing user speech;

calculating probabilities that particular pseudo noises will occur at locations of particular pseudo punctuation marks; and

finding most likely pseudo punctuation marks at locations of said pseudo noises marked in said step of marking pseudo noises, based on said probabilities calculated, and generating punctuation marks corresponding to said most likely pseudo punctuation marks.

4. An method for automatically generating punctuation marks in a continuous speech recognition system, comprising a step of recognizing user speech and converting said user speech into words, further comprising steps of:

generating a location indicating signal in response to operations of a user during dictation, said location indicating signal indicating locations in output results of said step of recognizing user speech;

calculating probability for each pseudo punctuation mark in a language model containing pseudo punctuation marks that said pseudo punctuation mark would occur in said output results of said step of recognizing user speech; and

finding most likely pseudo punctuation marks at locations indicated by said location indicating signal based on said probabilities calculated, and generating punctuation marks corresponding to said most likely pseudo punctuation marks.