Dynamic pronunciation support for Japanese and Chinese speech recognition training

• US 20040236581A1
• Filed: 05/01/2003
• Published: 11/25/2004
• Est. Priority Date: 05/01/2003
• Status: Abandoned Application

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Senju, Kazuhiro, Ju, Yun-Cheng, Hon, Hsiao-Wuen
• Application Number:
  US10/427,216
• Publication Number:
  US 20040236581A1
• Time in Patent Office: Days
• US Class Current: 704/276
• CPC Class Codes: G10L 15/063 Training
• CPC Class Codes: G10L 2015/0638 Interactive procedures

Abstract

A speech recognition training system for Kanji-based languages is provided. The system loads a pronunciation aid for each and every ideograph in the training speech, but does not in fact display an ideograph until the training system recognizes a pronunciation difficulty. Once a pronunciation difficulty is identified, the associated pronunciation aid (rubi) for the troubling ideograph is displayed.

Claims

1. A user interface module for speech recognition training, the module comprising:

a training text portion for displaying training text;

a communication channel coupleable to a speech recognition engine to receive an indication of pronunciation difficulty from the engine relative to a troubling word within the training text; and

a selectively displayed pronunciation aid disposed to aid pronunciation of the troubling word in response to the indication of pronunciation difficulty.

2. The module of claim 1, wherein the selectively displayed pronunciation aid is a rubi.

3. The module of claim 1, wherein the training text includes at least one ideograph.

4. The module of claim 3, wherein the training text is written in Chinese.

5. The module of claim 3, wherein the training text is written in Japanese.

6. The module of claim 1, wherein the pronunciation aid is displayed above of troubling word.

7. A method of training a speech recognition system, the method comprising:

loading a quantity of training text in the form of at least one ideographs;

displaying the training text without any pronunciation aid;

receiving a notification of pronunciation difficulty relative to a troubling word within the training text; and

selectively providing a pronunciation aid relative to the troubling word.

8. The method of claim 7, wherein the providing the pronunciation aid includes displaying a rubi to assist with pronunciation of the troubling word.

9. The method of claim 7, wherein the training text is written in Chinese.

10. The method of claim 7, wherein the training text is written in Japanese.

11. The method of claim 7, wherein the notification is received from a speech recognition engine.

12. The method of claim 7, wherein the pronunciation difficulty includes a pause.

13. The method of claim 7, wherein the pronunciation difficulty includes a mispronunciation.