Find an image's main colour.

What “dominant” means here

The dominant colour is the one the largest share of the image is closest to. Palettae works this out properly: once the palette is chosen, every pixel in the image is assigned to whichever palette colour it is nearest in perceptual terms, and the percentages come from those counts. That is why they add up to 100.

Plenty of tools report the share of pixels that fell inside the winning clusters and leave the rest unaccounted for, which is why their percentages often total something like 60%. Neither approach is wrong exactly, but only one of them answers the question people are actually asking.

Dominant is not the same as memorable

The biggest colour in a photograph is very often the least interesting thing in it.

A photograph of a red door on a grey street is mostly grey. If you rank strictly by pixel count you get five greys and no door. Palettae takes the dominant colour first and then weighs the remaining slots by how distinct and how saturated each candidate is — so the door survives, and its 4% is reported honestly rather than inflated.

What this is useful for

• Backgrounds and placeholders. The dominant colour is what to show behind an image while it loads.
• Cataloguing. Sorting a photo library or a product catalogue by main colour.
• Checking a brand photo. Whether the shot actually carries the brand colour, or only appears to.
• Moodboards. Seeing which images in a set agree with each other before you commit to a direction.

Sorting

The palette starts in dominance order. You can re-sort it light to dark, dark to light, or by hue — hue order puts the greys at the end rather than scattering them through the spectrum, since a grey’s hue is close to meaningless.