Image collation system, image collation method and computer program

• US 8,938,132 B2
• Filed: 12/14/2010
• Issued: 01/20/2015
• Est. Priority Date: 01/12/2010
• Status: Active Grant

• Current Assignee: NEC Corporation
• Original Assignee: NEC Corporation
• Inventors: Hamada, Yasushi
• Primary Examiner(s): BHATNAGAR, ANAND P
• Application Number:
  US13/521,436
• Publication Number:
  US 20120314904A1
• Time in Patent Office: 1,498 Days
• Field of Search: 382/100, 382/103, 382/108, 382/115, 382/116, 382/117, 382/118, 382/154, 382/181, 382/190, 382/195, 382/201, 382/203, 382/276, 382/278, 382/289, 382/294, 382/296, 708/422, 345619-689
• US Class Current: 382/294
• CPC Class Codes: G06T 2207/30201 Face
• CPC Class Codes: G06T 7/73 using feature-based methods
• CPC Class Codes: G06V 20/647 by matching two-dimensional images to three-dimensional objects
• CPC Class Codes: G06V 20/653 by matching three-dimensional models, e.g. conformal mapping of Riemann surfaces
• CPC Class Codes: G06V 30/2504 Coarse or fine approaches, e.g. resolution of ambiguities or multiscale approaches
• CPC Class Codes: G06V 40/16 Human faces, e.g. facial parts, sketches or expressions
• CPC Class Codes: G06V 40/161 Detection; Localisation; Normalisation

Abstract

An image collation system includes: a first direction estimating unit for estimating a first imaging direction of a reference object that matches an imaging direction of a collation target object by comparing global characteristics between an image of the collation target object and the three-dimensional data of the reference object; a second direction estimating unit for generating an image corresponding to the first imaging direction of the reference object, and estimating a second imaging direction of the reference object that matches the imaging direction of the collation target object by comparing local characteristics between the image of the collation target object and the generated image corresponding to the first imaging direction; and an image conformity determining unit for generating an image corresponding to the second imaging direction of the reference object, and determining whether the image of the collation target object matches the generated image corresponding to the second imaging direction.

Claims

1. An image collation system, comprising:

a three-dimensional data storing unit for storing three-dimensional data on at least a shape and a texture of a reference object;

a first direction estimating unit for estimating a first imaging direction of the reference object that matches an imaging direction of a collation target object by comparing global characteristics between an image of the collation target object and the three-dimensional data of the reference object stored in the three-dimensional data storing unit;

a second direction estimating unit for generating an image corresponding to the first imaging direction of the reference object based on the three-dimensional data, and estimating a second imaging direction of the reference object that matches the imaging direction of the collation target object by comparing local characteristics between the image of the collation target object and the generated image corresponding to the first imaging direction; and

an image conformity determining unit for generating an image corresponding to the second imaging direction of the reference object based on the three-dimensional data, and determining whether the collation target object corresponds to the reference object based on whether the image of the collation target object matches the generated image corresponding to the second imaging direction.

2. The image collation system according to claim 1, wherein

the second direction estimating unit extracts corresponding partial areas in the image of the collation target object and the image corresponding to the first imaging direction of the reference object, and determines the second imaging direction of the reference object so that the extracted corresponding partial areas overlap with each other.

3. The image collation system according to claim 2, wherein

the second direction estimating unit changes the imaging direction for each partial area of the image corresponding to the first imaging direction of the reference object, so that the partial area becomes most similar to each partial area of the image of the collation target object, and determines the partial area of the image of the collation target object of which similarity is highest, as the corresponding partial area.

4. The image collation system according to claim 3, wherein

the image conformity determining unit determines whether the image corresponding to the second imaging direction of the reference object matches the image of the collation target object, based on the result of comparing the global characteristics and local characteristics respectively between these images.

5. The image collation system according to claim 3, wherein

the collation target object and the reference object comprise human faces, and the local characteristics comprise image characteristics based on at least one of:

moles, spots, freckles, pores and bumps on the skin of the face.

6. The image collation system according to claim 2, wherein

the image conformity determining unit determines whether the image corresponding to the second imaging direction of the reference object matches the image of the collation target object, based on the result of comparing the global characteristics and local characteristics respectively between these images.

7. The image collation system according to claim 2, wherein

the collation target object and the reference object comprise human faces, and the local characteristics comprise image characteristics based on at least one of:

moles, spots, freckles, pores and bumps on the skin of the face.

8. The image collation system according to claim 1, wherein

the image conformity determining unit determines whether the image corresponding to the second imaging direction of the reference object matches the image of the collation target object, based on the result of comparing the global characteristics and local characteristics respectively between these images.

9. The image collation system according to claim 8, wherein

the collation target object and the reference object comprise human faces, and the local characteristics comprise image characteristics based on at least one of:

moles, spots, freckles, pores and bumps on the skin of the face.

10. The image collation system according to claim 1, wherein

the collation target object and the reference object comprise human faces, and the local characteristics comprise image characteristics based on at least one of:

moles, spots, freckles, pores and bumps on the skin of the face.

11. An image collation method comprising:

a first direction estimating of estimating a first imaging direction of a reference object that matches an imaging direction of a collation target object by comparing global characteristics between an image of the collation target object and three-dimensional data on at least a shape and a texture of the reference object stored in three-dimensional data storing means;

a second direction estimating of generating an image corresponding to the first imaging direction of the reference object based on the three-dimensional data and estimating a second imaging direction of the reference object that matches the imaging direction of the collation target object by comparing local characteristics between the image of the collation target object and the generated image corresponding to the first imaging direction; and

an image conformity determining of generating an image corresponding to the second imaging direction of the reference object based on the three-dimensional data, and determining whether the collation target object corresponds to the reference object based on whether the image of the collation target object matches the generated image corresponding to the second imaging direction.

12. A computer program stored on a non-transitory computer readable medium for causing a computer to execute the image collation method according to claim 11.