Recognition apparatus with function of displaying plural recognition candidates

• US 5,329,609 A
• Filed: 07/30/1991
• Issued: 07/12/1994
• Est. Priority Date: 07/31/1990
• Status: Expired due to Term

• Current Assignee: Fujitsu Limited
• Original Assignee: Fujitsu Limited
• Inventors: Loken-Kin, Kyung-Ho, Sanada, Toru, Kimura, Shinta
• Primary Examiner(s): Fleming, Michael R.
• Assistant Examiner(s): Doerrler, Michelle
• Application Number:
  US07/737,871
• Time in Patent Office: 1,078 Days
• Field of Search: 381/41-48, 395/2.85, 395/2.6, 395/2.44, 395/600, 382/30, 382/57, 364/200
• US Class Current: 704/251
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue

Abstract

A dictionary order sorter resorts character strings of recognition candidates stored in a high-ranking candidate memory in the order of distance into a dictionary order (character code order). Upon receipt of a sort termination signal a display controller displays the character strings of recognition candidates stored in the high-ranking candidate memory in the dictionary order and their ranking numbers in order on a display. Where an attribute-dependent sorter is provided in place of the dictionary order sorter, the character strings of recognition candidates stored in the distance order in the high-ranking candidate memory are sorted (grouped) according to attribute information of categories, such as parts of speech, concepts, etc., which are stored in a template memory for each of the recognition candidates and then displayed classified into groups. The categories of attribute information can arbitrarily be specified by the user. The user can easily select the correct recognition candidate from among the recognition candidates displayed rearranged in a fixed order of priority as described above.

Claims

1. A recognition apparatus for displaying a plurality of recognition candidates for an input voice or character and allowing a user to select a recognition candidate from among the displayed recognition candidates, comprising:

high-ranking candidate storage means for storing a predetermined number of high-ranking recognition candidates from the recognition candidates obtained in descending order of recognition degree;

sorting means for rearranging the order in which the recognition candidates are stored in said high-ranking candidate storage means in accordance with a fixed order of priority independent of the recognition degree; and

display means for displaying the recognition candidates in said high-ranking candidate storage means, which have been rearranged by said sorting means, in order.

2. A recognition apparatus according to claim 1, in which said sorting means rearranges the storing order in which the recognition candidates are stored in said high-ranking candidate storage means into the order of character codes.

3. A recognition apparatus according to claim 1, in which said sorting means rearranges the storing order in which the recognition candidates are stored in said high-ranking candidate storage means into alphabetical order by initials of the recognition candidates.

4. A recognition apparatus according to claim 1, in which said display means displays a recognition candidate having a high possibility of being correct and a recognition candidate having a low possibility of being correct in a visually distinguishable manner in displaying the recognition candidates in said highranking candidate storage means which have been rearranged by said sorting means.

5. A recognition apparatus according to claim 1, in which said display means displays the recognition candidates in said high-ranking candidate storage means which have been rearranged by said storing means to specify a recognition candidate a having the highest possibility of being correct.

6. A recognition apparatus for displaying a plurality of recognition candidates obtained by recognition process applied to an input voice or character and allowing a user to select a recognition result from among the displayed recognition candidates, comprising:

dictionary storage means for storing each of a plurality of objects of recognition which can be the recognition candidates together with attribute information, independent of recognition degree, representing a respective attribute; and

attribute-dependent displaying means for classifying the recognition candidates into groups according to said attribute information obtained from said dictionary storage means and displaying the recognition candidates for each classified group.

7. A recognition apparatus according to claim 6, in which said dictionary storage means stores a plurality of pieces of attribute information each representing an attribute of a category for each of the objects of recognition, and includes specifying means for allowing the user to specify said category of attribute information which is a candidate for classification by said attribute-dependent display means;

and

said attribute-dependent display means classifies the recognition candidates according to the attribute information of a category specified by the user through said specifying means which is attribute information corresponding to each of the recognition candidates obtained by referring to said dictionary storage means and displays the recognition candidates for each classified group.

8. A recognition apparatus according to claim 7, in which the categories include at least one a part of speech, a concept and the root of a word.

9. A recognition apparatus according to claim 7, in which said specifying means comprises a mouse.

10. A recognition apparatus according to claim 6, in which said attribute-dependent display means displays recognition candidates in a same group in descending order of recognition degree.

11. A recognition-apparatus according to claim 6, in which said attribute-dependent display means displays the recognition candidate in a visually distinguishable manner for each classified group.