Estimation of illumination chromaticity in automatic white balancing

1. A method, comprising:

calculating, via a white balancer, a candidate illumination chromaticity estimate for a current frame;

calculating, via a scene prior monitor, a scene prior chromaticity for the current frame as a weighted sum of a scene prior chromaticity from a previous frame and the candidate illumination chromaticity estimate, wherein the scene prior chromaticity from the previous frame comprises a frame confidence score that is determined as a sum of a single frame AWB confidence score and an illuminant change score;

calculating, via a scene invariant illumination chromaticity controller, a final illumination chromaticity for the current frame as a weighted sum of a scene illumination chromaticity for the current frame and the candidate illumination chromaticity estimate.

2. The method of claim 1 , wherein the scene prior chromaticity from the previous frame comprises a frame confidence score that is determined as a sum of a single frame AWB confidence score and an illuminant change score, wherein the single frame AWB confidence score identifies a likelihood that a single-frame based AWB algorithm estimate is correct.

3. The method of claim 2 , wherein the single frame AWB confidence score is based on an RGB histogram variability measure.

4. The method of claim 2 , wherein the illuminant change score is a likelihood that an illumination change occurs simultaneously with a scene change.

5. The method of claim 1 , wherein the scene prior chromaticity for the current frame is to constrain an image formation model to a most likely favorable solution.

6. The method of claim 1 , wherein the candidate illumination chromaticity estimate for the current frame is determined via a single frame based AWB algorithm.

7. The method of claim 1 , wherein the final illumination chromaticity corresponds to diagonal elements of a matrix A of white balancing gains.

8. The method of claim 1 , wherein the final illumination chromaticity is an estimate of white balancing gains in a temporal domain from a sequence of frames.

9. The method of claim 1 , wherein the scene prior chromaticity for the current frame is a temporally invariant representation for the current scene illuminant.

10. An apparatus, comprising:

a white balancer to calculate a candidate illumination chromaticity estimate for a current frame;

a scene prior monitor to calculate a scene prior chromaticity for the current frame as a weighted sum of a scene prior chromaticity from a previous frame and the candidate illumination chromaticity estimate, wherein the scene prior chromaticity from the previous frame comprises a frame confidence score that is determined as a weighted sum of a single frame AWB confidence score and an illuminant change score;

a scene invariant illumination chromaticity controller to calculate a final illumination chromaticity for the current frame as a weighted sum of scene illumination chromaticity for the current frame and the candidate illumination chromaticity estimate.

11. The apparatus of claim 10 , wherein the weighted scene prior chromaticity from the previous frame comprises a frame confidence score that is determined as a sum of a single frame AWB confidence score and an illuminant change score, wherein the single frame AWB confidence score identifies a likelihood that a single-frame based AWB algorithm estimate is correct.

12. The apparatus of claim 11 , wherein the single frame AWB confidence score is based on an RGB histogram variability measure.

13. The apparatus of claim 11 , wherein the illuminant confidence score is a likelihood that an illumination change occurs simultaneously with a scene change.

14. The apparatus of claim 10 , wherein the scene prior chromaticity for the current frame is to constrain an image formation model to a most likely favorable solution.

15. A system, comprising:

an image capture device;

a memory that is to store instructions and that is communicatively coupled to the image capture device; and

a processor communicatively coupled to the image capture device and the memory, wherein when the processor is to execute the instructions, the processor is to:

calculate a candidate illumination chromaticity estimate for a current frame;

calculate a scene prior chromaticity for the current frame as a weighted sum of a scene prior chromaticity from a previous frame and the candidate illumination chromaticity estimate, wherein the scene prior chromaticity from the previous frame comprises a frame confidence score that is determined as a sum of a single frame AWB confidence score and an illuminant change score; and

calculate a final illumination chromaticity for the current frame as a weighted sum of a scene illumination chromaticity for the current frame and the candidate illumination chromaticity estimate.

16. The system of claim 15 , wherein the scene prior chromaticity from the previous frame comprises a frame confidence score that is determined as a sum of a single frame AWB confidence score and an illuminant change score, wherein the single frame AWB confidence score identifies a likelihood that a single-frame based AWB algorithm estimate is correct.

17. The system of claim 15 , wherein the single frame AWB confidence score is based on an RGB histogram variability measure.

18. The system of claim 15 , wherein the illuminant change score is a likelihood that an illumination change occurs simultaneously with a scene change.

19. The system of claim 15 , wherein the scene prior chromaticity for the current frame is to constrain an image formation model to a most likely favorable solution.

20. The system of claim 15 , wherein the candidate illumination chromaticity estimate for the current frame is determined via a single frame based AWB algorithm.

21. The system of claim 15 , wherein the final illumination chromaticity corresponds to diagonal elements of a matrix A of white balancing gains.

22. The system of claim 15 , wherein the final illumination chromaticity is an estimate of white balancing gains in a temporal domain from a sequence of frames.