Method and apparatus for improving automatic white balance with scene information

1. A method for a digital signal processor for automatic white balancing of an image, comprising:

retrieving scene analysis of the image;

determining a scene classification map relating to light, a type of the scene relating to scene locality and a scene content map utilizing the scene analysis of the image;

performing scene adaptive white balance;

if overall scene category type is used, perform white balance by applying global R, G, B gains optimized for the global scene type, wherein the automated white balancing is performed more than once on at least a portion of the image utilizing scene information from a prior automatic white balancing; and

if scene segmentation map is used, perform locally adaptive white balance by applying an R, G, B gain map optimized for each scene content.

2. The method of claim 1 , wherein the step of determining the scene type comprises determining if the scene is at least one of an indoor scene, outdoor scene or a scene that includes a face.

3. The method of claim 1 , wherein the method is performed more than once on at least a portion of the image.

4. The method of claim 1 , wherein the automatic white balancing is performed with face prioritization.

5. The method of claim 1 , wherein the automatic white balancing is performed at least one of without scene mask, with scene masking and with locally adaptive automatic white balancing.

6. An apparatus for automatic white balancing of an image, comprising:

means for retrieving scene content information of the image;

determining a scene classification map relating to light, a type of the scene relating to scene locality and a scene content map utilizing the scene analysis of the image;

means for performing scene adaptive white balance, wherein the automated white balancing is performed more than once on at least a portion of the image utilizing scene information from a prior automatic white balancing;

If an overall scene category type is used, applying global R, G, B gains optimized for this scene category; and

if scene content map is used, applying locally adaptive white balance correction with a R, G, B gain map optimized for each scene content region.

7. The apparatus of claim 6 , wherein the means for determining the scene type comprises a means for determining if the scene is at least one of an indoor scene, outdoor scene or a scene that includes a face.

8. The apparatus of claim 6 , wherein the automatic white balancing is performed with face prioritization.

9. The apparatus of claim 6 , wherein the automatic white balancing is performed at least one of without scene masking, with scene masking, and with locally adaptive automatic white balancing.

10. A non-transitory computer readable medium comprising software that, when executed by a processor, causes the processor to perform a method for automatic white balancing of an image, the method comprising:

retrieving scene information of the image;

determining a scene classification map relating to light, a type of the scene relating to scene locality and a scene content map utilizing the scene analysis of the image;

performing scene adaptive white balance, wherein the automated white balancing is performed more than once on at least a portion of the image utilizing scene information from a prior automatic white balancing;

If an overall scene category type is used, applying global R, G, B gains optimized for this scene category; and

if scene content map is used, applying locally adaptive white balance correction with a R, G, B gain map optimized for each scene content region.

11. The computer readable medium of claim 10 , wherein the step of determining the scene type comprises determining if the scene is at least one of an indoor scene, outdoor scene or a scene that includes a face.

12. The computer readable medium of claim 10 , wherein the method is performed more than once on at least a portion of the image.

13. The computer readable medium of claim 10 , wherein the automatic white balancing is performed with face prioritization.

14. The computer readable medium of claim 10 , wherein the automatic white balancing is performed at least one of without scene masking, with scene masking, and with locally adaptive automatic white balancing.