Color quantization is the computational challenge of reducing the millions of colors in a typical photograph to a small representative palette, and several elegant algorithms have been developed to solve this problem. A standard 24-bit color image can contain up to 16.7 million distinct colors, but a useful palette typically contains only 5 to 16 colors that capture the essential color character of the image.

The median cut algorithm, introduced by Paul Heckbert in 1982, is one of the most widely used color quantization methods. It works by placing all pixels in a three-dimensional RGB color space and recursively subdividing the space along the axis of greatest range. First, the algorithm finds which color channel (red, green, or blue) has the widest spread across all pixels. It then sorts the pixels along that axis and splits the set at the median, creating two groups. This process repeats recursively, splitting the largest remaining group each time, until the desired number of color groups is reached. The representative color for each group is the average of all pixels within it. Median cut produces perceptually balanced palettes because it allocates more palette entries to color regions with more pixels.

K-means clustering offers an alternative approach borrowed from machine learning. The algorithm begins by randomly placing K color points (centroids) in the RGB color space. Each pixel is assigned to its nearest centroid, then each centroid is moved to the average position of its assigned pixels. This assignment-and-update cycle repeats until the centroids stabilize. K-means can produce more perceptually accurate palettes than median cut because it optimizes for minimum total distance between pixels and their representative colors, but it is sensitive to initial centroid placement and may converge to suboptimal solutions.

Color theory provides the framework for understanding why certain color combinations are harmonious. Complementary colors sit opposite each other on the color wheel (like blue and orange, or red and green) and create maximum contrast and visual energy when paired. Analogous colors sit adjacent on the wheel (like blue, blue-green, and green) and create harmonious, unified palettes. Triadic schemes use three colors equally spaced on the wheel for vibrant balance, while split-complementary schemes use a color plus the two colors adjacent to its complement for contrast with less tension. Understanding these relationships helps designers use extracted palettes effectively, recognizing whether an image naturally contains complementary, analogous, or other harmonic relationships that can inform design decisions.