Image processing method, image processing apparatus, and image processing program

• US 20070092132A1
• Filed: 11/02/2006
• Published: 04/26/2007
• Est. Priority Date: 07/26/2004
• Status: Active Grant

• Current Assignee: Matsushita Electric Industrial Company Limited (Panasonic Holdings Corporation)
• Original Assignee: Matsushita Electric Industrial Company Limited (Panasonic Holdings Corporation)
• Inventors: Nobori, Kunio, Sato, Satoshi, Takata, Kazutoyo
• Granted Patent: US 7,236,630 B2
• Time in Patent Office: Days
• US Class Current: 382/154
• CPC Class Codes: G06T 2207/20021 Dividing image into blocks, subimages or windows
• CPC Class Codes: G06T 5/20 using local operators
• CPC Class Codes: G06T 5/50 using two or more images, e.g. averaging or subtraction
• CPC Class Codes: G06T 5/94 based on local image properties, e.g. for local contrast enhancement
• CPC Class Codes: G06V 10/20 Image preprocessing

Abstract

An image region dividing section (105) partitions each of input images into a small regions, and a small region synthesized image generation section (106) generates a base image of each small region, which is an image including only diffuse reflection, from each input image. An image interpolation section (107) generates by interpolation a base image of a small region of which base image cannot be generated. A linearized image generation section (103) generates a linearized image, which is an image in an ideal state under given lighting condition, using the base image of each small region.

Claims

1. An image processing method comprising:

a first step of obtaining a plurality of input images obtained by shooting a same object;

a second step of dividing each of the obtained input images into a plurality of small regions;

a third step of generating a base image, which includes only diffuse reflection, of each of the divided small regions using the input images; and

a fourth step of generating a base image of a small region having been impossible to generate a base image by interpolation processing using a computed value for base image generation in neighbor small regions, the small region and the neighbor small regions being of the divided small regions,

wherein a linearized image of the object under given lighting condition is generated using the generated base images.

2. The method of claim 1, further comprising the step of:

performing image processing using the generated linearized image.

3. The method of claim 2, wherein

the image processing includes at least one of:

region classification based on an optical characteristic of an object;

object identification;

object 3D position/shape estimation.

4. The method of claim 1, wherein

the plurality of images are obtained by shooting the object while changing lighting condition.

5. The method of claim 4, wherein

a light source is controlled so that a shadow cast on the object is different in position from each other in the plurality of images.

6. The method of claim 1, wherein

the plurality of images are obtained by shooting the object when variation in lighting condition is detected.

7. The method of claim 6, wherein

when a shadow cast on the object varies, it is detected that the lighting condition varies.

15. The method of claim 1, wherein

in the second step,

a size of the small regions is determined based on at least one of an upper limit value of light expanse that can be regarded as a parallel light source, a distance between a light source and the object, a distance between a camera and the object, and a camera parameter.

16. The method of claim 1, wherein

a size of the small region of which base image has been ungenerated is changed and the third step is executed again, instead of the interpolation in the fourth step.

17. The method of claim 1, wherein

a position of the small region of which base image has been ungenerated is changed and the third step is executed again, instead of the interpolation in the fourth step.

18. The method of claim 1, wherein

the plurality of input images are shot under environment where a plurality of light sources exist.

19. The method of claim 1, wherein

the plurality of input images are shot in a home.

20. The method of claim 1, wherein

the method is executed by a robot that performs work in a home.

21. The method of claim 1, wherein

the object is a face of a person.

22. The method of claim 1, wherein

the method is utilized for biometrics authentication.

23. The method of claim 2, wherein

the image processing is processing for estimating position of a protrusion or a dimple included in the object.

8-14. -14. (canceled)

24. An image processing apparatus comprising:

an image input section that obtains a plurality of images obtained by shooting a same object;

an image region dividing section that divides each of the input images obtained by the image input section into a plurality of small regions;

a small region synthesized image generation section that generates a base image, which includes only diffuse reflection, of each of the small regions divided by the image region dividing section using the input images; and

an image interpolation section that generates a base image of a small region having been impossible to generate a base image in the small region synthesized image generation section by interpolation processing using a computed value for base image generation in neighbor regions near to the small region, the small region and the neighbor small regions being of the partitioned small regions,

wherein a linearized image of the object under given lighting condition is generated using the generated base images.

25. An image processing program on computer-readable medium for allowing a computer to execute image processing, which allows a computer to execute the steps of:

dividing a plurality of input images obtained by shooting a same object into a plurality of small regions;

generating a base image, which includes only diffuse reflection, of each of the divided small regions using the input images;

generating a base image of a small region having been impossible to generate a base image by interpolation processing using a computed value for base image generation in neighbor small regions near to the small region, the small region and the neighbor small region being of the partitioned small regions; and

generating a linearized image of the object under given lighting condition using the generated base images.