Input/output apparatus based on voice recognition, and method thereof

• US 8,478,600 B2
• Filed: 09/11/2006
• Issued: 07/02/2013
• Est. Priority Date: 11/11/2005
• Status: Expired due to Fees

• Current Assignee: Electronics and Telecommunications Research Institute
• Original Assignee: Electronics and Telecommunications Research Institute
• Inventors: Cho, Kwan-Hyun, Han, Mun-Sung, Park, Jun-Seok, Jung, Young-Giu
• Primary Examiner(s): ALBERTALLI, BRIAN LOUIS
• Application Number:
  US12/093,091
• Publication Number:
  US 20080288260A1
• Time in Patent Office: 2,486 Days
• Field of Search: None
• US Class Current: 704/275
• CPC Class Codes: G06F 3/011 Arrangements for interaction with the human body, e.g. for user immersion in virtual reality blind teaching G09B21/00
• CPC Class Codes: G06F 3/04842 Selection of displayed objects or displayed text elements G06F3/0482 takes precedence
• CPC Class Codes: G06F 3/16 Sound input; Sound output speech processing G10L

Abstract

Provided is an input/output apparatus based on voice recognition, and a method thereof. An object of the apparatus is to improve a user interface by making pointing input and command execution such as application program control possible according to a voice command of a user possible based on a voice recognition technology without individual pointing input device such as a mouse and a touch pad, and a method thereof. The apparatus includes: a voice recognizer for recognizing a voice command inputted from outside; a pointing controller for calculating a pointing location on a screen which corresponds to a voice recognition result transmitted from the voice recognizer; a displayer for displaying a screen; and a command controller for processing diverse commands related to a current pointing location.

Claims

1. An input/output apparatus based on voice recognition, comprising:

a voice recognizing means for recognizing a voice command inputted from outside;

a pointing control means for calculating a pointing location on a screen which corresponds to a voice recognition result transmitted from the voice recognizing means;

a displaying means for displaying a screen; and

a command control means for processing diverse commands related to a current pointing location;

wherein the voice recognizing means performs analog-to-digital (A/D) transformation on a user voice corresponding to the voice command, extracts a feature vector based on the A/D transformation user voice, and performs a pattern matching for the feature vector by using a recognition model parameter;

wherein the pointing control means performs mapping on a voice recognition command corresponds to the voice recognition result to a voice recognition candidate command;

wherein the displaying means displays the screen by duplicating a region corresponding to the calculated pointing location on a prior screen;

wherein the screen is divided into blocks, and the blocks are divided into sub-blocks;

wherein neighbor blocks in the blocks identically include parts of the sub-blocks, and the neighbor blocks are duplicated by the parts of the sub-blocks.

2. The apparatus as recited in claim 1, wherein the command control means processes event commands including click, double-clicks and scrolling.

3. The apparatus as recited in claim 1, wherein the command control means processes execution commands for controlling an application program and setting up of a system.

4. The apparatus as recited in claim 1, wherein the pointing control means calculates the pointing location by dividing the screen into a plurality of regions having a predetermined size.

5. The apparatus as recited in claim 4, wherein the pointing control means calculates the pointing location by re-dividing each region in multiple levels for minute pointing.

7. The apparatus as recited in claim 1, wherein the feature vector includes a Linear Predictive Coefficients (LPC), a Mel-frequency Cepstral Coefficients (MFCC), and a Perceptual Linear Prediction (PLP).

8. The apparatus as recited in claim 1, wherein the recognition model parameter is obtained by using a Dynamic Time Warping (DTW), a Neural Network, and a Hidden Markov Model (HMM).

6. An input/output method based on voice recognition, comprising the steps of:

a) recognizing a voice command inputted from outside;

b) calculating a pointing location on a screen corresponding to the recognized voice command;

c) displaying the calculated pointing location to be identified; and

d) executing diverse commands related to the pointing location;

wherein the step of a), performs analog-to-digital (A/D) transformation on a user voice corresponding to the voice command, extracts a feature vector based on the A/D transformation user voice, and performs a pattern matching for the feature vector by using a recognition model parameter;

wherein the step of b), performs mapping on a voice recognition command corresponds to the voice recognition result to a voice recognition candidate command;

wherein the step of c), displays the screen by duplicating a region corresponding to the calculated pointing location on a prior screen;

wherein the screen is divided into blocks, and the blocks are divided into sub-blocks;

wherein neighbor blocks in the blocks identically include parts of the sub-blocks, and the neighbor blocks are duplicated by the parts of the sub-blocks.

9. The method as recited in claim 6, wherein event commands for click, double-click and scrolling are processed in the step d).

10. The method as recited in claim 6, wherein execution commands for controlling an application program or setting up of a system are processed in the step d).

11. The method as recited in claim 6, wherein the pointing location is calculated by dividing the screen into a plurality of regions having predetermined size in the step b).

12. The method as recited in claim 11, wherein in the step b), the regions are re-divided in multiple levels and a minute pointing location is calculated in a sequence of the voice commands.

13. The method as recited in claim 6, wherein the feature vector includes a Linear Predictive Coefficients (LPC), a Mel-frequency Cepstral Coefficients (MFCC), and a Perceptual Linear Prediction (PLP).

14. The method as recited in claim 6, wherein the recognition model parameter is obtained by using a Dynamic Time Warping (DTW), a Neural Network, and a Hidden Markov Model (HMM).