System and method for computer recognition and interpretation of arbitrary spoken-characters

• US 20050203742A1
• Filed: 03/08/2005
• Published: 09/15/2005
• Est. Priority Date: 03/09/2004
• Status: Active Grant

• Current Assignee: Ashwin Rao
• Original Assignee: Ashwin Rao
• Inventors: Rao, Ashwin
• Granted Patent: US 7,865,363 B2
• Time in Patent Office: Days
• US Class Current: 704/254
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue

Abstract

For improving the accuracy of a speech recognition system, for the specific task of character recognition, the character that is to be recognized or interpreted, is spoken in accordance with this invention, using simple coding-schemes that result in acoustic renderings that are well separated in the acoustic space. Furthermore, to relax the constraint on users for memorizing multiple words associated with multiple characters, in one embodiment, the coding-schemes are defined based on a combination of the given character and its neighboring/succeeding character(s).

Claims

1. A method for recognizing spoken characters, the method comprising:

a) for each spoken character to be recognized there exists a specific existing predetermined sequence of one or more subsequent characters to be also spoken; and

b) by using a predetermined number of said spoken characters in said sequence then determining the specific spoken character that initiated said sequence.

2. The method of claim 1 further including displaying text corresponding to the recognition of the spoken character.

3. The method of claim 1 wherein the characters spoken are letters and the existing predetermined sequence is a language'"'"'s alphabet.

4. The method of claim 1 wherein the characters spoken are numbers and the existing predetermined sequence is the language'"'"'s number system.

5. The method of claim 2 wherein the text of the character displayed also generates an audio pronunciation thereby giving a user of the system audio feedback.

6. The method of claim 2 wherein a string of said displayed text characters comprises a word.

7. The method of claim 6 wherein a string of said displayed words comprises a text message.

8. The method of claim 7 wherein the text message may be wirelessly communicated to another system and displayed on said another system.

9. The method of claim 1 wherein errors in a general purpose speech recognition system are corrected by a user.

10. The method of claim 1 wherein an acoustic model or language model is improved by a user.