Speech recognition method computer readable medium and apparatus for recognizing geographical names using weight information

• US 7,624,011 B2
• Filed: 12/08/2004
• Issued: 11/24/2009
• Est. Priority Date: 12/12/2003
• Status: Expired due to Fees

• Current Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Original Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Inventors: Fukada, Toshiaki
• Primary Examiner(s): Smits; Talivaldis I
• Assistant Examiner(s): PULLIAS, JESSE SCOTT
• Application Number:
  US11/006,628
• Publication Number:
  US 20050131699A1
• Time in Patent Office: 1,812 Days
• Field of Search: 704/240, 340/426.19
• US Class Current: 704/240
• CPC Class Codes: G10L 15/00 Speech recognition G10L17/00 takes precedence
• 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 15/08 Speech classification or search
• CPC Class Codes: G10L 15/10 using distance or distortion measures between unknown speech and reference templates
• CPC Class Codes: G10L 15/14 using statistical models, e.g. Hidden Markov Models [HMMs] G10L15/18 takes precedence
• CPC Class Codes: G10L 15/18 using natural language modelling
• CPC Class Codes: G10L 15/1815 Semantic context, e.g. disambiguation of the recognition hypotheses based on word meaning
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence
• CPC Class Codes: G10L 15/28 Constructional details of speech recognition systems
• CPC Class Codes: G10L 2015/228 of application context

Abstract

A speech recognition apparatus and method of this invention manage previously input frequencies of occurrence for respective geographical names to be recognized (202), update the probability of occurrence of the geographical name to be recognized of interest on the basis of the frequency of occurrence of that geographical name, and those of geographical names to be recognized located within a predetermined region including the position of the geographical name of interest using a table (114) that describes correspondence between the geographical names to be recognized and their positions, and perform this update process for respective geographical names to be recognized (203).

Claims

1. A method for performing speech recognition of geographical names using a weighting factor associated with respective geographical names to be speech recognized, the method comprising:

a speech recognition step of performing speech recognition in which likelihood values of entered speech for each recognition target geographical name are calculated, wherein each likelihood value is weighted by a weighting factor corresponding to a recognition target geographical name;

a determination step of determining a geographical name having a maximum likelihood value as a speech recognition result;

a frequency-of-occurrence update step of updating frequency of occurrence information of the geographical name obtained by the determination step, wherein the frequency of occurrence information indicates the number of times that the geographical name has been determined as a speech recognition result in the determination step;

a first obtaining step of obtaining a position of the geographical name from a table that describes correspondence between geographical names to be speech recognized and their geographical positions;

a region determination step of determining a region centered on the position of the geographical name;

an extraction step of extracting geographical names to be speech recognized located within the region on the basis of the table;

a frequency-of-occurrence obtaining step of obtaining the frequency of occurrence information of the extracted geographical names;

a second obtaining step of obtaining positions of the extracted geographical names;

a distance factor determination step of determining distance factors based on distances between the position of the geographical name obtained in the first obtaining step and the positions of the extracted geographical names obtained in the second obtaining step; and

a weighting factor update step of updating the weighting factor associated with the geographical name determined as the speech recognition result in the determination step by using the updated frequency of occurrence information of the geographical name obtained in the frequency-of-occurrence update step, the frequency of occurrence information of the extracted geographical names obtained in the frequency-of-occurrence obtaining step, and the distance factors determined in the distance factor determination step,

wherein the updated weighting factor is used in the next speech recognition step.

2. A computer-readable storage medium storing a computer program for performing speech recognition of geographical names using a weighting factor associated with respective geographical names to be speech recognized, the computer program comprising:

code for a speech recognition step of performing speech recognition in which likelihood values of entered speech for each recognition target geographical name are calculated, wherein each likelihood value is weighted by a weighting factor corresponding to a recognition target geographical name;

code for a determination step of determining a geographical name having a maximum likelihood value as a speech recognition result;

code for a frequency-of-occurrence update step of updating frequency of occurrence information of the geographical name obtained by the code for the determination step, wherein the frequency of occurrence information indicates the number of times that the geographical name has been determined as a speech recognition result in the code for the determination step;

code for a first obtaining step of obtaining a position of the geographical name from a table that describes correspondence between geographical names to be speech recognized and their geographical positions;

code for a region determination step of determining a region centered on the position of the geographical name;

code for an extraction step of extracting geographical names to be speech recognized located within a region on the basis of the table;

code for a frequency-of-occurrence obtaining step of obtaining the frequency of occurrence information of the extracted geographical names;

code for a second obtaining step of obtaining positions of the extracted geographical names;

code for a distance factor determination step of determining distance factors based on distances between the position of the geographical name obtained in the code for the first obtaining step and the positions of the extracted geographical names obtained in the code for the second obtaining step; and

code for a weighting factor update step of updating the weighting factor associated with the geographical name determined as the speech recognition result in the code for the determination step by using the updated frequency of occurrence information of the geographical name obtained in the code for the frequency-of-occurrence update step, the frequency of occurrence information of the extracted geographical names obtained in the code for the frequency-of-occurrence obtaining step, and the distance factors determined in the code for the distance factor determination step,

wherein the updated weighting factor is used in the code for the next speech recognition step.

3. A speech recognition apparatus for performing speech recognition of geographical names using a weighting factor associated with respective geographical names to be speech recognized, the apparatus comprising:

a speech recognition means for performing speech recognition in which likelihood values of entered speech for each recognition target geographical name are calculated, wherein each likelihood value is weighted by a weighting factor corresponding to a recognition target geographical name;

a determination means for determining a geographical name having a maximum likelihood value as a speech recognition result;

a frequency-of-occurrence update means for updating frequency of occurrence information of the geographical name obtained by the determination means, wherein the frequency of occurrence information indicates the number of times that the geographical name has been determined as a speech recognition result in the determination means;

a first obtaining means for obtaining a position of the geographical name from a table that describes correspondence between geographical names to be speech recognized and their geographical positions;

a region determination means for determining a region centered on the position of the geographical name;

an extracting means for extracting geographical names to be speech recognized located within a region on the basis of the table;

a frequency-of-occurrence obtaining means for obtaining the frequency of occurrence information of the extracted geographical names;

a second obtaining means for obtaining positions of the extracted geographical names;

a distance factor determination means for determining distance factors based on distances between the position of the geographical name obtained by the first obtaining means and the positions of the extracted geographical names obtained by the second obtaining means; and

a weighting factor update means for updating the weighting factor associated with the geographical name determined as the speech recognition result by the determination means by using the updated frequency of occurrence information of the geographical name obtained by the frequency-of-occurrence update means, the frequency of occurrence information of the extracted geographical names obtained by the frequency-of-occurrence obtaining means, and the distance factors determined by the distance factor determination step,

wherein the updated weighting factor is used by the next speech recognition means.