Methods and apparatus for recognized word registration in accordance with speech recognition

• US 6,983,248 B1
• Filed: 09/07/2000
• Issued: 01/03/2006
• Est. Priority Date: 09/10/1999
• Status: Expired due to Term

• Current Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Original Assignee: International Business Machines Corporation
• Inventors: Tahara, Yoshinori, Uda, Tetsuya, Tomoda, Daisuke, Ichikawa, Norikazu
• Primary Examiner(s): Chawan, Vijay
• Assistant Examiner(s): Opsasnick, Michael N.
• Application Number:
  US09/656,964
• Time in Patent Office: 1,944 Days
• Field of Search: 704/255, 704/231, 704/251, 704/254
• US Class Current: 704/255
• CPC Class Codes: G10L 15/06 Creation of reference templates; Training of speech recognition systems, e.g. adaptation to the characteristics of the speaker's voice G10L15/14 takes precedence
• CPC Class Codes: G10L 15/187 Phonemic context, e.g. pronunciation rules, phonotactical constraints or phoneme n-grams
• CPC Class Codes: G10L 2015/0631 Creating reference templates; Clustering

Abstract

A sounds-like spelling group is generated in which scores are used for word descriptions. A user selects a correct sounds-like spelling from a generated sounds-like spelling group, and a group of base forms is generated with scores provided for the selected sounds-like spelling and the reading of the word. In this fashion, a base form having a score that exceeds a reference value is registered without a voice recording being required. During the speech recognition process as performed by a user, when a predetermined error state is detected, e.g., when the number of recognition errors exceeds N, a panel is output requesting that the user register the inscription, the sounds-like spelling and the pronunciation inscription for the word and that the user record a corresponding pronunciation. In accordance with the pronunciation provided by the user, the base form is obtained and is re-registered in the speech recognition dictionary.

Claims

1. A recognized word registration method, for a speech recognition apparatus that includes a display screen and a voice input device, comprising the steps of:

obtaining a word inscription specified by a user;

searching a word dictionary to obtain a sounds-like spelling corresponding to said word inscription;

searching a pronunciation dictionary to obtain a base form corresponding to said sounds-like spelling that has been obtained; and

registering said base form in a speech recognition dictionary.

2. A recognized word registration method, for a speech recognition apparatus that includes a display screen and a voice input device, comprising the steps of:

obtaining a word inscription specified by a user;

searching a word dictionary to obtain a plurality of sounds-like spellings that correspond to said word inscription and sounds-like spelling scores that correspond to said sounds-like spellings;

displaying said plurality of sounds-like spellings for said user;

obtaining said sounds-like spelling that is selected by said user from among said plurality of sounds-like spellings;

searching a pronunciation dictionary to obtain a base form and a pronunciation score corresponding to said sounds-like spelling that has been obtained;

determining whether said pronunciation score exceeds a predetermined threshold value; and

registering said base form in a speech recognition dictionary when said pronunciation score exceeds said predetermined threshold value.

3. A recognized word registration method, for a speech recognition apparatus that includes a display screen and a voice input device, comprising the steps of:

determining whether first voice information obtained for a user'"'"'s voice matches a predetermined condition;

displaying on said display screen, when said voice information matches said predetermined condition, a speech recognition wizard panel that includes a new word input field and a sounds-like spelling input field;

obtaining a new word and a sounds-like spelling that are entered in said speech recognition wizard panel;

obtaining second voice information based on said user'"'"'s pronunciation provided for said new word and said sounds-like spelling;

employing said second voice information, said new word and said sounds-like spelling to specifically describe a base form; and

adding said base form to a speech recognition dictionary.

4. A recognized word registration method, for a speech recognition apparatus that includes a display screen and a voice input device, comprising:

an initial registration step, including;

obtaining a word inscription specified by a user,

searching a word dictionary to obtain a plurality of sounds-like spellings that correspond to said word inscription and sounds-like spelling scores that correspond to said sounds-like spellings,

displaying said plurality of sounds-like spellings for said user, obtaining said sounds-like spelling that is selected by said user from among said plurality of sounds-like spellings,

searching a pronunciation dictionary to obtain a base form and a pronunciation score corresponding to said sounds-like spelling that has been obtained,

determining whether said pronunciation score exceeds a predetermined threshold value, and

registering said base form in a speech recognition dictionary when said pronunciation score exceeds said predetermined threshold value, and

a registration step at the speech recognition time, including;

determining whether first voice information obtained for a user'"'"'s voice matches a predetermined condition,

displaying on said display screen, when said voice information matches said predetermined condition, a speech recognition wizard panel that includes a new word input field and a sounds-like spelling input field,

obtaining a new word and a sounds-like spelling that are entered in said speech recognition wizard panel,

obtaining second voice information based on said user'"'"'s pronunciation provided for said new word and said sounds-like spelling,

employing said second voice information, said new word and said sounds-like spelling to specifically describe a second base form, and

adding said second base form to a speech recognition dictionary.

5. A speech recognition apparatus, which includes a display screen and a voice input device, comprising:

a recognized word registration unit for obtaining a word inscription specified by a user;

a sounds-like spelling generator for searching a word dictionary to obtain a sounds-like spelling corresponding to said word inscription;

a base form generator for searching a pronunciation dictionary to obtain a base form corresponding to said sounds-like spelling that has been obtained; and

a speech recognition dictionary in which said base form is registered.

6. A speech recognition apparatus, which includes a display screen and a voice input device, comprising:

a recognized word registration unit for obtaining a word inscription specified by a user;

a sounds-like spelling generator for searching a word dictionary to obtain a plurality of sounds-like spellings that correspond to said word inscription and sounds-like spelling scores that correspond to said sounds-like spellings, and for obtaining said sounds-like spelling that is selected by said user from among said plurality of sounds-like spellings on said display screen;

a base form generator for searching a pronunciation dictionary to obtain a base form and a pronunciation score corresponding to said sounds-like spelling that has been obtained; and

a speech recognition dictionary in which said base form is registered when said pronunciation score exceeds a predetermined threshold value.

7. A speech recognition apparatus, which includes a display screen and a voice input device, comprising:

a recognized word registration unit for determining whether first voice information obtained for a user'"'"'s voice matches a predetermined condition;

a speech recognition wizard for displaying on said display screen, when said voice information matches said predetermined condition, a speech recognition wizard panel that includes a new word input field and a sounds-like spelling input field;

a voice input unit for obtaining second voice information based on said user'"'"'s pronunciation provided for a new word and a sounds-like spelling that are entered in said speech recognition wizard panel;

a base form generator for employing said second voice information, said new word and said sounds-like spelling to specifically describe a base form; and

a speech recognition dictionary to which said base form is added.

8. A speech recognition apparatus comprising:

a display screen;

a voice input unit for entering voice information generated by a user'"'"'s voice;

a speech recognition engine for recognizing said voice information;

a recognized word registration unit for obtaining a word inscription specified by a user;

a sounds-like spelling generator for searching a word dictionary to obtain a plurality of sounds-like spellings that correspond to said word inscription and sounds-like spelling scores that correspond to said sounds-like spellings, and for, when one of said plurality of sounds-like spellings is selected by said user, obtaining said sounds-like spelling that is selected;

a base form generator for searching a pronunciation dictionary to obtain a base form and a pronunciation score corresponding to said sounds-like spelling that has been obtained; and

a speech recognition dictionary in which a base form is registered when said pronunciation score exceeds said predetermined threshold value;

wherein said speech recognition engine determines whether first voice information obtained for a user'"'"'s voice matches a predetermined condition, and activates, when said voice information matches said predetermined condition, a speech recognition wizard panel that includes a new word input field and a sounds-like spelling input field;

wherein said sounds-like spelling generator obtains second voice information based on said user'"'"'s pronunciation provided for a new word and a sounds-like spelling that are entered in said speech recognition wizard panel;

wherein said base form generator employs said second voice information, said new word and said sounds-like spelling to specifically describe a second base form; and

wherein said second base form is stored in said speech recognition dictionary.

9. A storage medium on which is stored a recognized word registration program that is to be executed by a speech recognition apparatus that includes a display screen and a voice input device, said recognized word registration program comprising:

program code for instructing said speech recognition apparatus to obtain a word inscription specified by a user;

program code for instructing said speech recognition apparatus to search a word dictionary to obtain a sounds-like spelling corresponding to said word inscription;

program code for instructing said speech recognition apparatus to search a pronunciation dictionary to obtain a base form corresponding to said sounds-like spelling that has been obtained; and

program code for instructing said speech recognition apparatus to register said base form in a speech recognition dictionary.

10. A storage medium on which is stored a recognized word registration program that is to be executed by a speech recognition apparatus that includes a display screen and a voice input device, said recognized word registration program comprising:

program code for instructing said speech recognition apparatus to obtain a word inscription specified by a user;

program code for instructing said speech recognition apparatus to search a word dictionary to obtain a plurality of sounds-like spellings that correspond to said word inscription and sounds-like spelling scores that correspond to said sounds-like spellings;

program code for instructing said speech recognition apparatus to display said plurality of sounds-like spellings for said user;

program code for instructing said speech recognition apparatus to obtain said sounds-like spelling that is selected by said user from among said plurality of sounds-like spellings;

program code for instructing said speech recognition apparatus to search a pronunciation dictionary to obtain abase form and a pronunciation score corresponding to said sounds-like spelling that has been obtained;

program code for instructing said speech recognition apparatus to determine whether said pronunciation score exceeds a predetermined threshold value; and

program code for instructing said speech recognition apparatus to register said base form in a speech recognition dictionary when said pronunciation score exceeds said predetermined threshold value.

11. A storage medium on which is stored a speech recognition process program that is to be executed by a speech recognition apparatus that includes a display screen and a voice input device, said speech recognition process program comprising:

program code for instructing said speech recognition apparatus to determine whether first voice information obtained for a user'"'"'s voice matches a predetermined condition;

program code for instructing said speech recognition apparatus to display on said display screen, when said voice information matches said predetermined condition, a speech recognition wizard panel that includes a new word input field and a sounds-like spelling input field;

program code for instructing said speech recognition apparatus to obtain a new word and a sounds-like spelling that are entered in said speech recognition wizard panel;

program code for instructing said speech recognition apparatus to obtain second voice information based on said user'"'"'s pronunciation provided for said new word and said sounds-like spelling;

program code for instructing said speech recognition apparatus to employ said second voice information, said new word and said sounds-like spelling to specifically describe a base form; and

program code for instructing said speech recognition apparatus to add said base form to a speech recognition dictionary.

12. A storage medium on which is stored a speech recognition process program that is to be executed by a speech recognition apparatus that includes a display screen and a voice input device, said speech recognition process program comprising:

program code for instructing said speech recognition apparatus to obtain a word inscription specified by a user;

program code for instructing said speech recognition apparatus to search a word dictionary to obtain a plurality of sounds-like spellings that correspond to said word inscription and sounds-like spelling scores that correspond to said sounds-like spellings;

program code for instructing said speech recognition apparatus to display said plurality of sounds-like spellings for said user;

program code for instructing said speech recognition apparatus to obtain said sounds-like spelling that is selected by said user from among said plurality of sounds-like spellings;

program code for instructing said speech recognition apparatus to search a pronunciation dictionary to obtain a base form and a pronunciation score corresponding to said sounds-like spelling that has been obtained;

program code for instructing said speech recognition apparatus to determine whether said pronunciation score exceeds a predetermined threshold value;

program code for instructing said speech recognition apparatus to register said base form in a speech recognition dictionary when said pronunciation score exceeds said predetermined threshold value;

program code for instructing said speech recognition apparatus to determine whether first voice information obtained for a user'"'"'s voice matches a predetermined condition;

program code for instructing said speech recognition apparatus to display on said display screen, when said voice information matches said predetermined condition, a speech recognition wizard panel that includes a new word input field and a sounds-like spelling input field;

program code for instructing said speech recognition apparatus to obtain a new word and a sounds-like spelling that are entered in said speech recognition wizard panel;

program code for instructing said speech recognition apparatus to obtain second voice information based on said user'"'"'s pronunciation provided for said new word and said sounds-like spelling;

program code for instructing said speech recognition apparatus to employ said second voice information, said new word and said sounds-like spelling to specifically describe a second base form; and

program code for instructing said speech recognition apparatus to add said second base form to a speech recognition dictionary.