GAZE TRACKING APPARATUS AND METHOD USING DIFFERENCE IMAGE ENTROPY

• US 20090284608A1
• Filed: 02/09/2009
• Published: 11/19/2009
• Est. Priority Date: 05/15/2008
• Status: Active Grant

• Current Assignee: Sungkyunkwan University Foundation For Corporate Collaboration (SungKyunKwan University)
• Original Assignee: Sungkyunkwan University Foundation For Corporate Collaboration (SungKyunKwan University)
• Inventors: Hong, Kwang Seok, Lee, Kue Bum, Jeon, Jong Bae, Yoon, Jun Ho
• Granted Patent: US 8,274,578 B2
• Time in Patent Office: Days
• US Class Current: 348/207.990
• CPC Class Codes: G06F 3/013 Eye tracking input arrangements G06F3/015 takes precedence
• CPC Class Codes: G06T 7/248 involving reference images or patches
• CPC Class Codes: G06V 40/20 Movements or behaviour, e.g. gesture recognition recognition of facial expressions G06V40/16

Abstract

In accordance with one aspect of the present invention, there is provided a gaze tracking apparatus using difference image entropy, comprising: image detection module for detecting user'"'"'s eye region image from image of the user taken by picture-taking apparatus which is connected to the gaze tracking apparatus; image selection module for computing entropy value of difference image between the eye region image and a plurality of reference images respectively indicating each gazing direction and selecting reference image with least difference image entropy value among the plurality of reference images; and gaze tracking module for recognizing user'"'"'s gazing direction by regarding gazing direction that the selected reference image indicates as user'"'"'s gazing direction.

Claims

1. A gaze tracking apparatus using difference image entropy, comprising:

image detection module for detecting user'"'"'s eye region image from image of the user taken by picture-taking apparatus which is connected to the gaze tracking apparatus;

image selection module for computing entropy value of difference image between the eye region image and a plurality of reference images respectively indicating each gazing direction and selecting reference image with least difference image entropy value among the plurality of reference images; and

gaze tracking module for recognizing user'"'"'s gazing direction by regarding gazing direction that the selected reference image indicates as user'"'"'s gazing direction.

2. The apparatus of claim 1, wherein the image detection module comprises:

face region detection unit for detecting the user'"'"'s face region image from the image of the user taken by the picture-taking apparatus; and

eye region detection unit for detecting the user'"'"'s eye region image from the user'"'"'s face region image.

3. The apparatus of claim 2, wherein the image detection module uses at least one of Haar-like features, AdaBoost learning algorithm, Template matching, and Entropy method.

4. The apparatus of claim 1, wherein the image selection module computes difference image between the eye region image and the plurality of reference images using Equation below.

D(x, y)=Iinput(x, y)−

I(x, y)

Where D (x, y) denotes pixel value of difference image between the eye region image and the reference images, Iinput (x, y) denotes pixel value of the eye region image and I (x, y) denotes pixel value of the reference image.

5. The apparatus of claim 4, wherein the image selection module computes entropy value of difference image between the eye region image and the reference images by Equation below.

6. The apparatus of claim 1,

wherein the image selection module computes entropy value of difference image between the eye region image and average images respectively representing each gazing direction instead of the plurality of reference images and selects average image with least difference image entropy value among the average images,

wherein the average images are obtained by taking an average of pixel values of the reference images indicating each gazing direction, and

wherein the number of average images is equal to the number of the gazing directions.

7. A gaze tracking method using difference image entropy, comprising:

detecting user'"'"'s eye region image from inputted user'"'"'s image;

computing entropy value of difference image between the eye region image and a plurality of reference images respectively indicating each gazing direction and selecting reference image with least difference image entropy value among the plurality of reference images; and

recognizing user'"'"'s gazing direction by regarding gazing direction that the selected reference image indicates as user'"'"'s gazing direction.

8. The method of claim 7, wherein said detecting user'"'"'s eye region image comprises:

detecting the user'"'"'s face region image from the image of the user; and

detecting the user'"'"'s eye region image from the user'"'"'s face region image.

9. The method of claim 8, wherein said detecting user'"'"'s eye region image uses at least one of Haar-like features, AdaBoost learning algorithm, Template matching, and Entropy method.

10. The method of claim 7,

wherein the difference image between the eye region image and the plurality of reference images is computed using Equation below.

D(x, y)=Iinput(x, y)−

I(x, y)

Where D (x, y) denotes pixel value of difference image between the eye region image and the reference images, Iinput (x, y) denotes pixel value of the eye region image and I (x, y) denotes pixel value of the reference image.

11. The method of claim 10, wherein entropy value of difference image between the eye region image and the reference images is computed by Equation below.

12. The method of claim 7,

wherein entropy value of difference image is computed between the eye region image and average images respectively representing each gazing direction, instead of the plurality of reference images, and average image with least difference image entropy value among the average images is selected,

wherein the average images are obtained by taking an average of pixel values of the reference images indicating each gazing direction, and

wherein the number of average images is equal to the number of the gazing directions.

13. An information processing apparatus comprising:

picture-taking module for taking images of user;

gaze tracking module for detecting eye region image from the image taken by the picture-taking module, computing entropy value of difference image between the eye region image and a plurality of reference images respectively indicating each gazing direction, and recognizing user'"'"'s gazing direction by regarding gazing direction corresponding to reference image with least difference image entropy value as the user'"'"'s gazing direction; and

control module for controlling to display a plurality of menu which user can select and to perform process designed for menu corresponding to the recognized gazing direction.

14. The apparatus of claim 13, wherein the gaze tracking module computes difference image between the eye region image and the plurality of reference images using Equation below.

D(x, y)=Iinput(x, y)−

I(x, y)

Where D (x, y) denotes pixel value of difference image between the eye region image and the reference images, Iinput (x, y) denotes pixel value of the eye region image and I (x, y) denotes pixel value of the reference image.

15. The apparatus of claim 14, wherein the gaze tracking module computes entropy value of difference image between the eye region image and the reference images by Equation below.