Method and apparatus for estimating light source

1. A light source estimating apparatus comprising:

a brightness-color correlation calculator which calculates brightness-color correlation data from an image of an object generated from at least one image frame;

a color balance calculator which calculates a color balance of an entire region of the image and a color balance of a neutral gray region of the image having an average color of gray;

a characteristic data generating unit which generates characteristic data of the image based on the brightness-color correlation data, the color balance of the entire region and the color balance of the neutral gray region; and

an identifier which determines a type of a light source in the image, based on the characteristic data of the image.

2. The light source estimating apparatus of claim 1 , further comprising a texture analyzing unit which calculates a color texture characteristic indicating color diversity of the image,

wherein the characteristic data generating unit generates the characteristic data of the image based on the color texture characteristic of the image.

3. The light source estimating apparatus of claim 1 , further comprising a brightness irregularity analyzer which specifies a constant brightness region having constant brightness in the image,

wherein the brightness-color correlation calculator calculates the brightness-color correlation data in the constant brightness region in the image.

4. The light source estimating apparatus of claim 1 , wherein the identifier comprises:

a learning unit which learns a determination basis for extracting the type of the light source from the characteristic data of the image, based on learning data comprising learned characteristic data of a plurality of model images and types of light sources corresponding to the plurality of model images, respectively; and

an identifying unit which determines the type of the light source corresponding to the characteristic data of the image, based on a result of the learning by the learning unit.

5. The light source estimating apparatus of claim 4 , wherein the identifying unit determines the type of the light source by comparing the characteristic data of the image with the learned characteristic data of the plurality of model images, selects learned characteristic data of a model image most similar to the characteristic data of the image, and determines a type of a light source corresponding to the selected learned characteristic data of the model image as the type of the light source in the image.

6. The light source estimating apparatus of claim 5 , wherein the characteristic data of the image and the type of the light source in the image becomes one of the learned characteristic data of the plurality of model images constituting the learning data which is used for determining a type of a light source in another image of another object.

7. The light source estimating apparatus of claim 1 , wherein the identifier determines the type of the light source by applying different weights to the brightness-color correlation data and the color balance of the entire region of the image based on a number of colors detected in the image.

8. The light source estimating apparatus of claim 1 , wherein the brightness-color correlation calculator calculates the brightness-color correlation data from the image of the object generated from a plurality of image frames obtained sequentially with respect to the same object.

9. The light source estimating apparatus of claim 1 , wherein the image of the object generated from at least one image frame from which the brightness-color correlation calculator calculates the brightness-color correlation data is image data which is compressed or reduced from original image data of the object obtained by an imaging device.

10. The light source estimating apparatus of claim 1 , wherein the type of the light source comprises a color of the light source.

11. A method of estimating a light source, the method comprising:

calculating brightness-color correlation data from an image of an object generated from at least one image frame;

calculating a color balance of an entire region of the image and a color balance of a natural gray region of the image having an average color or gray;

generating characteristic data of the image based on the brightness-color correlation data, the color balance of the entire region and the color balance of the neutral gray region; and

determining a type of a light source in the image, based on the characteristic data of the image.

12. The method of claim 11 , further comprising calculating a color texture characteristic indicating color diversity of the image,

wherein the characteristic data is generated further based on the color texture characteristic.

13. The method of claim 11 , further comprising specifying a constant brightness region having constant brightness in the image,

wherein the brightness-color correlation data is calculated with respect to the constant brightness region.

14. The method of claim 11 , wherein the determining the type of the light source in the image comprises:

learning a determination basis for extracting the type of the light source from the characteristic data of the image, based on learning data comprising learned characteristic data of a plurality of model images and types of light sources corresponding to the plurality of images, respectively; and

determining the type of the light source corresponding to the characteristic data of the image, based on a result of the learning.

15. The method of claim 14 , wherein the determining the type of the light source comprises:

comparing the characteristic data of the image with the learned characteristic data of the plurality of model images;

selecting learned characteristic data of a model image most similar to the characteristic data of the image; and

determining a type of a light source corresponding to the selected learned characteristic data of the model image as the type of the light source in the image.

16. The method of claim 15 , wherein the characteristic data of the image and the type of the light source in the image becomes one of the learned characteristic data of the plurality of model images constituting the learning data which is used for determining a type of a light source in another image of another object.

17. The method of claim 11 , wherein the determining the type of the light source comprises applying different weights to the brightness-color correlation data and the color balance of the entire region of the image based on a number of colors detected in the image.

18. The method of claim 11 , wherein the brightness-color correlation data is calculated from the image of the object generated from a plurality of image frames obtained sequentially with respect to the same object.

19. The method of claim 11 , wherein the image of the object generated from at least one image frame from which the brightness-color correlation data is calculated is image data which is compressed or reduced from original image data of the object obtained by an imaging device.

20. The method of claim 11 , wherein the type of the light source comprises a color of the light source.