Standard pattern comparing system for eliminating duplicative data entries for different applications program dictionaries, especially suitable for use in voice recognition systems

• US 5,355,433 A
• Filed: 03/18/1991
• Issued: 10/11/1994
• Est. Priority Date: 03/26/1990
• Status: Expired due to Term

• Current Assignee: Ricoh Company Limited
• Original Assignee: Ricoh Company Limited
• Inventors: Yasuda, Seigou, Grennan, Peter
• Primary Examiner(s): Fleming, Michael R.
• Assistant Examiner(s): Doerrler, Michelle
• Application Number:
  US07/670,789
• Time in Patent Office: 1,303 Days
• Field of Search: 381/41-46, 395/2.52
• US Class Current: 704/243
• CPC Class Codes: G10L 15/063 Training
• CPC Class Codes: G10L 2015/0635 updating or merging of old and new templates; Mean values; Weighting

Abstract

A standard pattern comparing system recognizes a data by comparing the data with a standard pattern registered in a dictionary, a plurality of kinds of data being supplied to the standard pattern comparing system. The standard pattern comparing system has a plurality of dictionaries including a plurality of normal dictionaries and one master dictionary. All standard patterns stored in each normal dictionary and some of the standard patterns stored in the master dictionary corresponding to one kind of data. There being no duplicate standard pattern included in more than two dictionaries, and each standard pattern included in the master dictionary being commonly used for at least more than two normal dictionaries to correspond to an arbitrary kind of data. The present invention is especially useful for a voice recognition system in which a speaker dependent voice recognizer is used for a personal computer which manages a plurality of application programs. A user of the voice recognizer does not have to register some duplicate words commonly used for at least more than two application programs because of the master dictionary.

Claims

1. A standard pattern comparing system to which a plurality of kinds of data are supplied and which recognizes data by comparing the data with standard patterns registered in a dictionary, the system comprising:

a) a memory which stores a plurality of dictionaries, the dictionaries including;

1) a plurality of normal dictionaries, and

2) one master dictionary;

wherein;

1) all standard patterns stored in each normal dictionary correspond to one kind of data;

2) there are no duplicate standard-patterns included in at least two normal dictionaries, and

3) each standard pattern included in the master dictionary is commonly used for at least two normal dictionaries to correspond to an arbitrary kind of data;

b) selection means, coupled to the memory, for selecting a corresponding one of the normal dictionaries and the master dictionary from the memory, based on the data;

c) comparing means, responsive to the selection means, for comparing the data with one of the standard patterns stored in the corresponding one of the normal dictionaries and the master dictionary;

d) output means, responsive to the comparing means, for outputting a comparison result for recognizing the data; and

e) analyzing means for analyzing the feature of the data so as to generate a feature pattern corresponding to the feature.

2. A standard pattern comparing system to which a plurality of kinds of data are supplied and which recognizes data by comparing the data with standard patterns registered in a dictionary, the system comprising:

a) a memory which stores a plurality of dictionaries, the dictionaries including;

1) a plurality of normal dictionaries; and

2) one master dictionary;

wherein;

1) all standard patterns stored in each normal dictionary correspond to one kind of data;

2) there are no duplicate standard patterns included in at least two normal dictionaries; and

3) each standard pattern included in the master dictionary is commonly used for at least two normal dictionaries to correspond to an arbitrary kind of data;

b) selection means, coupled to the memory, for selecting a corresponding one of the normal dictionaries and the master dictionary from the memory, based on the data;

c) comparing means, responsive to the selection means, for comparing the data with one of the standard patterns stored in the corresponding one of the normal dictionaries and the master dictionary; and

d) output means, responsive to the comparing means, for outputting a comparison result for recognizing the data.

3. The system of claim 2, wherein:

a) the standard pattern comparing system is especially suitable for use in a voice recognition system;

b) the data includes a voice input to the voice recognition system; and

c) the voice recognition system includes;

1) a voice recognizer operating in conjunction with the standard pattern comparing system;

2) a personal computer for controlling operations of the voice recognizer, the voice being a kind of data corresponding to an application program used in the personal computer; and

3) an external storage device connected to the personal computer.

5. A standard pattern comparing system according to claim 3, wherein said voice recognizer further comprises:

voice analyzing means for analyzing the voice input;

comparing means, responsive to the analyzing means, for comparing an analyzing result of the analyzing means with the standard patterns;

judging means, responsive to the comparing means, for judging whether or not the result of the analyzing means corresponds to the standard pattern; and

plotter means, responsive to the judging means, for plotting a judging result of the judging means as a recognition result.

4. A method for writing out a plurality of dictionaries in a standard pattern comparing system to which a plurality of kinds of data are supplied and which recognizes data by comparing the data with standard patterns registered in a dictionary, the method comprising the steps of:

a) writing out a plurality of basic dictionaries, each basic dictionary being written out by gathering all standard patterns required for an arbitrary kind of data; and

b) writing out (1) a plurality of normal dictionaries and (2) one master dictionary, by extracting from each basic dictionary a duplicate standard pattern included in at least two basic dictionaries;

wherein;

all standard patterns stored in each normal dictionary correspond to one kind of data;

there are no duplicate standard patterns included in at least two normal dictionaries; and

each standard pattern included in the master dictionary is commonly used for at least two normal dictionaries to correspond to an arbitrary kind of data.

6. The method of claim 4, wherein:

a) the writing out steps include using a standard pattern comparing system;

b) the standard pattern comparing system is especially suitable for use in a voice recognition system;

c) the data includes a voice input to the voice recognition system; and

d) the voice recognition system includes;

1) a voice recognizer operating in conjunction with the standard pattern comparing system;

2) a personal computer for controlling operations of the voice recognizer, the voice being a kind of data corresponding to an application program used in the personal computer; and

3) an external storage device connected to the personal computer.

7. The method of claim 6, wherein the voice recognizer further includes:

voice analyzing means for analyzing a voice input;

comparing means, responsive to the analyzing means, for comparing an analyzing result of the analyzing means with the standard patterns;

judging means, responsive to the comparing means, for judging whether or not the analyzing result of the analyzing means corresponds to the standard pattern; and

plotter means, responsive to the judging means, for plotting as a recognition result, a judging result of the judging means.

8. A method for writing out a plurality of dictionaries in a standard pattern comparing system to which a plurality of kinds of data are supplied and which recognizes data by comparing the data with standard patterns registered in a dictionary, the method comprising the steps of:

A) writing out a plurality of basic dictionaries, each basic dictionary being written out by gathering all standard patterns required for an arbitrary kind of data; and

B) writing out (1) a plurality of normal dictionaries and (2) one master dictionary, by extracting from each basic dictionary a duplicate standard pattern included in at least two basic dictionaries;

wherein;

a) all standard patterns stored in each normal dictionary correspond to one kind of data;

b) there are no duplicate standard patterns included in at least two normal dictionaries;

c) each standard pattern included in the master dictionary is commonly used for at least two normal dictionaries to correspond to an arbitrary kind of data;

d) the master and normal dictionary further include;

1) a voice dictionary for specifying a feature of a pronunciation of a word of a voice input; and

2) a string dictionary for specifying the spelling of the word thereof; and

e) the standard pattern of the basic dictionary includes voice and string data corresponding to the voice and string dictionaries.

9. The method of claim 8, wherein a duplicate standard pattern is extracted by a step of:

comparing each string data of the standard pattern of the basic dictionary with the string dictionaries of the master and normal dictionaries.

10. The method of claim 8, wherein a duplicate standard pattern is extracted by a step of:

comparing each voice data of the standard pattern of the basic dictionary with the voice dictionaries of the master and normal dictionaries.