Reducing computational complexity in determining an illuminant of a scene

摘要

In an embodiment, computational complexity of estimating the actual illuminant of a scene is reduced by examining only a subset of the pixel values generated for a received image frame. In another embodiment, number of rotations of color values is minimized by selecting an area which contains the color cue values of a color in an original/unrotated coordinate space and has boundaries which parallel the axis of the original coordinate space, and rotating a color value only if the color value is within the selected area. In another embodiment, such an area is used in conjunction with a histogram-based approach to determine the actual illuminant.

主权项

1. A method of determining a scene illuminant illuminating a scene of interest, said method comprising:
receiving an image frame and an information, wherein said image frame contains a plurality of color values representing said scene of interest and said information comprises a corresponding plurality of hue values associated with each of a set of colors; forming an area, wherein all hue values of said set of colors lie within said area; checking whether each of said plurality of color values is in said area; and determining whether a first color value matches a first color only if said first color value is in said area, wherein said first color value is deemed to match said first color only if any of the corresponding hue values equals said first color value, wherein the scene illuminant is determined based on a result of said determining for said plurality of color values including said first color value, wherein said area is a subset of pixels of said image frame that is selected according to a pattern.