Menu

Image Color Picker

Image Color Picker: How It Works

Picking a colour from an image reads the exact pixel value under the cursor. That is useful for matching a brand colour, building a palette from a photograph, or identifying a shade you cannot name — and there are a few reasons the value you get may not be the value you expected.

What you get

FormatExampleUse in
HEX#3B82F6CSS, design tools
RGBrgb(59 130 246)Code, programmatic use
HSLhsl(217 91% 60%)Building tints and shades
CMYK (approx.)76 47 0 4Print reference only

CMYK values from a screen colour are an approximation. Screens emit light and mix additively; print reflects it and mixes subtractively. A vivid screen colour may simply be outside the printable gamut, which is why a design can look correct on a monitor and dull on paper. For print work, specify colours from a printed reference or a Pantone book rather than converting from a screenshot.

Why the value may surprise you

For an authoritative brand colour, use the value from the brand guidelines rather than sampled pixels. Sampling is for approximation and inspiration.

Building a palette from an image

  1. Sample the dominant colour first — the largest area, not the most striking.
  2. Sample two or three supporting colours from different regions.
  3. Pick one accent from the most saturated small area.
  4. Derive neutrals by desaturating the dominant hue rather than using pure grey, which makes the set feel coherent.
  5. Check every text-and-background pairing for contrast before using them.

That last step is the one usually skipped. A palette pulled from a photograph is frequently beautiful and unreadable — colours of similar lightness look distinct while failing contrast badly.

Averaging a region

Sampling a single pixel from a textured or noisy area gives an unrepresentative value. Averaging a small region — a 5 × 5 or 11 × 11 block — returns something closer to the colour you perceive, which is what you usually want from a photograph. Single-pixel sampling is right only for flat, uncompressed graphics.

Accessibility check

Once you have a pair of colours, verify contrast: 4.5:1 for body text and 3:1 for large text or interface components under WCAG AA. And never rely on colour alone to convey meaning, since around 8% of men have some form of colour vision deficiency.

Sampling here happens entirely in your browser — the image is read by the page and never uploaded.

Frequently Asked Questions

Why doesn't the sampled colour match the brand's official value?
Compression artefacts, anti-aliasing near edges and colour profile conversion all shift pixel values. Sample from the middle of a large flat area, and use the brand's published value when accuracy matters.
Are the CMYK values accurate for printing?
They are approximations. Screens mix light additively and print mixes ink subtractively, and vivid screen colours are often outside the printable gamut. For print, specify from a printed reference rather than a screen conversion.
Why do I get a different colour near an edge?
Anti-aliasing blends edge pixels between the two adjoining colours, so an edge pixel is neither. Always sample from well inside a solid region.
Should I sample one pixel or an area?
An area for photographs and textured images, since a single pixel is often unrepresentative. A single pixel only for flat, uncompressed graphics where every pixel in a region is identical.
How do I build a usable palette from a photo?
Take the dominant colour, two or three supporting ones, and a single saturated accent. Derive neutrals by desaturating the dominant hue rather than using pure grey — and check every text pairing for contrast before committing.
Is my image uploaded?
No. The image is read by the page in your browser and never transmitted, so private photographs and unreleased design work stay on your device.

Related Image Tools

Browse all Image tools →