Recursive technique for 3-D histogram enhancement of color images

Histogram modification is often used for contrast enhancement of color images, usually by separately modifying the color components in some appropriately chosen color space. Despite its computational efficiency, independent component processing is seldom able to exploit the full RGB gamut. While joint processing of color components can generally utilize the full display gamut, such approaches generally produce greater computational complexity and execution times. Introduced here is the `histogram decimation' algorithm for enhancement by joint processing of the color components. The algorithm recursively partitions the 3-D histogram of an image and independently processes the subdivided volumes of color space to modify their color distributions. Histogram decimation achieves the gamut usage of other joint processing techniques and provides much greater computational efficiency. The algorithm and its characteristics are presented, followed by the results of processing several images and measuring the uniformity of their histograms using a novel technique.