Image recognition method and apparatus

• US 20060126900A1
• Filed: 02/08/2006
• Published: 06/15/2006
• Est. Priority Date: 12/25/1998
• Status: Active Grant

• Current Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Original Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Inventors: Yamauchi, Yasunobu, Morishita, Akira, Mihara, Isao, Doi, Miwako
• Granted Patent: US 7,394,918 B2
• Time in Patent Office: Days
• US Class Current: 382/103
• CPC Class Codes: G06T 2207/30241 Trajectory
• CPC Class Codes: G06T 7/246 using feature-based methods, e.g. the tracking of corners or segments
• CPC Class Codes: G06V 40/20 Movements or behaviour, e.g. gesture recognition recognition of facial expressions G06V40/16

Abstract

An image recognition apparatus has an image capture section for capturing a range image, an image deformation section for deforming the range image captured by the image capture section, and a recognition section for recognizing the presence/absence of three-dimensional motion of an object by comparing the deformed image obtained by the image deformation section, and a new range image captured by the image capture section.

Claims

1. (canceled)

2. An image recognition method comprising:

obtaining a deformed image by deforming a captured range image; and

recognizing the presence/absence of three-dimensional motion of an object in the range image by comparing the obtained deformed image with a newly captured range image, and recognizing a series of motions recognized from each of a series of a plurality of range images.

3. A method according to claim 2, further comprising predicting motion of the object on the basis of the series of recognized motions.

5. A method according to claim 2, further comprising predicting motion of the object on the basis of a feature amount of the object extracted from the captured range image.

6. A method according to claim 2, further comprising predicting motion of the object on the basis of a feature amount of the object extracted from the captured range image, and the series of recognized motions.

8. A method according to claim 2, further comprising compressing a range image captured by image capture unit on the basis of the recognized motion of the object.

4. (canceled)

7. (canceled)

9. (canceled)

10. An image recognition apparatus comprising:

an image capture unit configured to capture a range image;

an image deformation unit configured to deform the range image captured by said image capture unit;

a first recognition unit configured to recognize three-dimensional motion of an object by comparing a deformed image obtained by said image deformation unit and a new range image captured by said image capture unit; and

a second recognition unit configured to recognize a series of motions recognized from each of a series of a plurality of range images by said first recognition unit.

13. An apparatus according to claim 10, further comprising:

a prediction unit configured to predict motion of the object on the basis of the series of motions recognized by said second recognition unit.

17. An apparatus according to claim 10, further comprising:

a feature amount extraction unit configured to extract a feature amount of the object from the range image captured by said image capture unit; and

a prediction unit configured to predict motion of the object on the basis of the feature amount extracted by said feature amount extraction unit.

20. An apparatus according to claim 10, further comprising:

a feature amount extraction unit configured to extract a feature amount of the object from the range image captured by said image capture unit; and

a prediction unit configured to predict motion of the object on the basis of the feature amount extracted by said feature amount extraction unit, and the series of motions recognized by said second recognition unit.

24. An apparatus according to claim 13, further comprising:

an image compression unit configured to compress the range image captured by said image capture unit on the basis of the recognized motion of the object.

11. An image recognition apparatus comprising:

an image capture unit configured to capture a range image;

a storage unit configured to store the range image captured by said image capture unit;

an image deformation unit configured to deform a designated range image of the range image captured by said image capture unit and/or the range image stored in said storage unit; and

a recognition unit configured to recognize the presence/absence of three-dimensional motion of an object by comparing one of a deformed image obtained by said image deformation unit and the range image stored in said storage unit, and a new range image captured by said image capture unit.

14. An apparatus according to claim 11, further comprising:

a prediction unit configured to predict motion of the object on the basis of the series of motions recognized by said second recognition unit.

18. An apparatus according to claim 11, further comprising:

a feature amount extraction unit configured to extract a feature amount of the object from the range image captured by said image capture unit; and

a prediction unit configured to predict motion of the object on the basis of the feature amount extracted by said feature amount extraction unit.

21. An apparatus according to claim 11, further comprising:

a feature amount extraction unit configured to extract a feature amount of the object from the range image captured by said image capture unit; and

a prediction unit configured to predict motion of the object on the basis of the feature amount extracted by said feature amount extraction unit, and the series of motions recognized by said second recognition unit.

25. An apparatus according to claim 11, further comprising:

an image compression unit configured to compress the range image captured by said image capture unit on the basis of the recognized motion of the object.

12. An image recognition apparatus comprising:

an image capture unit configured to capture a range image;

a storage unit configured to store the range image captured by said image capture unit;

an image deformation unit configured to deform a designated range image of the range image captured by said image capture unit and/or the range image stored in said storage unit;

a first recognition unit configured to recognize three-dimensional motion of an object by comparing one of a deformed image obtained by said image deformation unit and the range image stored in said storage unit, and a new range image captured by said image capture unit; and

a second recognition units configured to recognize a series of motions recognized from each of a series of a plurality of range images by said first recognition unit.

15. An apparatus according to claim 12, further comprising:

a prediction unit configured to predict motion of the object on the basis of the series of motions recognized by said second recognition unit.

19. An apparatus according to claim 12, further comprising:

a feature amount extraction unit configured to extract a feature amount of the object from the range image captured by said image capture unit; and

a prediction unit configured to predict motion of the object on the basis of the feature amount extracted by said feature amount extraction unit.

22. An apparatus according to claim 12, further comprising:

a feature amount extraction unit configured to extract a feature amount of the object from the range image captured by said image capture unit; and

a prediction unit configured to predict motion of the object on the basis of the feature amount extracted by said feature amount extraction unit, and the series of motions recognized by said second recognition unit.

26. An apparatus according to claim 12, further comprising:

image compression unit configured to compress the range image captured by said image capture unit on the basis of the recognized motion of the object.

16. (canceled)

23. (canceled)

27. (canceled)

28. An article of manufacture comprised of a computer-usable medium having computer-readable program code means that implements computer-readable program code means for recognizing an image, comprising:

computer-readable program code means for making a computer capture a range image;

computer-readable program code means for making the computer deform the range image captured by the image capture means;

computer-readable program code means for making the computer recognize the presence/absence of three-dimensional motion of an object by comparing a deformed image obtained by the image deformation means and a new range image captured by the image capture means; and

computer-readable program code means for making the computer recognize a series of motions recognized from a series of a plurality of range images by the recognition means.