Color palettes from images: methods, color systems and tools for practical application

Anyone who works with color knows this moment: A photo, a sketch, or a painting perfectly captures the mood a new project needs. But how do you translate that mood into a coherent palette that can then be reliably reproduced on a book cover, a wall color, a logo, or an interior design?

The answer rarely lies in a single eye color, nor in an app's automatic filter. It lies in a combination of observation, a suitable color system, and a tool that combines both.

This article summarizes the methods currently used in studios, workshops, and design offices to extract colors from a sample. It explains the systems in which these colors can be reliably recorded and what needs to be considered when transferring them to print, wall paint, or digital screens. The aim is to provide a factual overview that is helpful for both beginners and experienced practitioners.

Three ways to get a palette from an image

1. Manual pipette

The classic method involves selecting individual pixels from a photo using an eyedropper tool. It's precise, but slow and highly dependent on the resolution and color profile of the original image. A JPG from a smartphone camera already contains several processing steps, including white balance, contrast adjustment, and sharpening. Therefore, an eyedropper tool displays the color interpreted by the sensor and software, not necessarily the actual color as seen in the scene.

The pipette is useful when focusing on a specific area in the original image: the eye of a person portrayed, a particular brushstroke on a canvas, a specific strip of fabric in a mixed-media collage. It is less suitable for capturing entire moods.

2. Dominant colors via K-means clustering

If a compact palette of five, ten, or fifteen colors is to be extracted from an image, K-means clustering become the standard. The algorithm groups the pixels in the color space (usually LAB or OKLCH) into a predefined number of classes and outputs their mean values. The result is a statistically representative palette that favors larger, contiguous color areas.

K-means is fast and easy to understand, but it has a well-known weakness: it treats every pixel the same. A blue sky that fills two-thirds of the image dominates the result, even if the actual subject is a bright red coat in the foreground.

3. Saliency-based extraction

Newer tools therefore separate the image foreground and background before clustering. Saliency maps weight pixels according to how strongly they attract the eye: edges, contrasts, faces, and very colorful areas are favored. In a photograph of a market scene, the observer first perceives the five most important colors of the foreground (for example, the fruit, the vendors' clothing, the awnings), then ten colors of the background (cobblestones, walls, sky).

For mood boards, branding research, and preparing a wall color, the saliency method is often the most helpful because it answers the question "Which colors shape my visual impression?" more directly than pure frequency.

Which color scheme is appropriate for which occasion?

A HEX notation like #7A6A55 is unambiguous on the screen. In the physical world, however, it says little. A carpenter, a painter, or an industrial designer needs a bridge to a system connected to sample boards, pigments, or paints.

The most important systems at a glance:

• RAL Classic / RAL Design: The most widely used system in Germany for paints, coatings, industrial paints, and wall finishes. RAL Classic contains 213 colors, RAL Design approximately 1,825. Mandatory in the construction and interior design sectors.
• NCS (Natural Colour System): A Scandinavian tradition, strongly used in architectural theory. It describes colors based on their black, white, and hue components, as well as their tint. Very useful for systematically comparing tints and saturations.
• Munsell: An academic system with hue, value, and chroma. Standard in color research, soil science, restoration, and the US school system.
• HKS: Widely used, especially in German-language printing, and accepted as a special color by many publishers and printers.
• Pantone equivalent: International standard for spot colors in printing and fashion. Anyone dealing with a printing company or textile manufacturer will rarely encounter it without using it.
• Federal Standard 595: Originally developed for the US military, today often a reference for historically accurate model colors, vintage restoration and special coatings.
• OKLCH: A modern, perceptually equidistant color space that is increasingly becoming the standard in web design because it produces uniform brightness and saturation steps.

A reliable conversion from screen values ​​to one of these systems is only possible using a color difference algorithm. ΔE 2000 has become established as a reliable standard: values ​​below 1 are considered practically indistinguishable, values ​​between 1 and 2.3 are just barely perceptible, and values ​​above 2.3 represent a visible deviation. Anyone choosing a wall paint or varnish based on a screen reference should know the ΔE value of their chosen color.

From screen to reality: Print, gamut and light context

A saturated cyan or a vibrant lime green can appear brilliant on a modern display, but in four-color printing (CMYK) can turn into a pale, unintentionally shifted tone. The reason is the limited color gamut of the printing process. Therefore, honest palette planning considers, before the first proof print, which values ​​fall outside the CMYK gamut and replaces them early with achievable alternatives.

The second pitfall is lighting. A color that looks perfect under D65 (daylight, around 6500 Kelvin) can appear different under D50 (printing standard, around 5000 Kelvin). Under incandescent light (standard A), cool tones shift significantly. Metameric pairs are colors that appear identical under one light source but diverge under another. Anyone planning a restaurant, museum, or apartment should experiment with at least two lighting scenarios.

Harmonies as a starting point

An extracted palette is rarely the final one. It often serves as the main color from which a composition is then developed. Classical harmonies help in the construction:

• Complementary: Main color plus its opposite on the color wheel. Maximum contrast, good for accents.
• Analog: Three neighboring colors on the color wheel. Calm, harmonious, often used for landscape painting and branding.
• Triadic: Three colors equally spaced. Vibrant, balanced.
• Split-complementary: A primary color with its two complementary neighbors. Soft contrast.
• Tetradic and square: Four colors for more complex compositions, where one acts as an anchor and three as accents.

Working in the OKLCH space results in consistent brightness and saturation steps. This is particularly helpful in UI design because it ensures that buttons, backgrounds, and text have a clear and consistent relationship to each other.

Tools that combine these methods in practice

There are now a number of programs that combine color extraction, cross-system matching, and harmonics in a single step. Shademix , a free tool that runs without registration and transmits nothing from the user's device, is a prime example of this category. It works in a browser, as an iPhone, iPad, and Mac app, as well as a browser extension and Figmaplugin. The individual components demonstrate how the methods described above interact.

Web app and mobile apps. In the main application, you upload an image and receive fifteen dominant colors in one step: five from the foreground using the saliency method and ten from the background. Each color value can then be converted to HEX, RGB, HSL, HSV, CMYK, LAB, OKLCH, and a Pantone equivalent. The cross-system search finds the closest RAL, NCS, Munsell, HKS, or Federal Standard 595 equivalent via ΔE 2000.

Browser extension. This Chrome extension provides a color picker tool that lets you select colors directly from any open tab. Inspiration sources like Pinterest, Behance, or online shops don't need to be downloaded first to examine a color. The extension saves your recent selections locally and doesn't transmit any data to external servers.

Figma plugin. Those working in UI or branding can integrate this feature directly into their workflow. The Figma plugin offers image extraction with fifteen colors per image, cross-system matching according to RAL, NCS, Munsell, HKS, Federal Standard 595 and Pantone equivalents, as well as OKLCH harmonies. The colors used can be transferred from a selection of layers with a single click. The plugin can convert a JSON file exported from Shademix into native Figma color styles.

Alongside these building blocks, the project maintains comprehensive, multilingual documentation on shademix.app, including a German manual as a PDF and individual pages for each tool. Anyone wishing to try one of the described methods will find a suitable starting point there.

A practical workflow from photo to palette

The following order has proven successful in many studios:

1. Image selection: Choose the rawest possible original image. JPEGs with heavy post-processing distort the color palette. A RAW or TIFF version is ideal; a well-exposed JPG with an sRGB profile is often sufficient.
2. Extraction: Saliency-based method, fifteen colors or fewer, depending on complexity.
3. Define the main color: From the extracted selection, highlight the color that conveys the mood. This will serve as the anchor point for the subsequent harmony.
4. Cross-System-Match: For each retained color, find the nearest match in the system needed in the subsequent process (RAL for wall paint, HKS or Pantone for printing, NCS for architectural specifications).
5. Gamut check: When the result goes to print, run a CMYK preview and replace any out-of-gamut values.
6. Deriving harmonies: Generating complementary, analogous, or triadic colors from the main color and using them as accents.
7. Light test: Test the palette under at least two different lighting conditions. Screen simulation or physical hanging of a sample.
8. Export: In a format that the next stage of the process understands. Designer tools often use ASE, Procreate its own .swatches, printers usually use HEX lists with assigned spot colors, and painters use a RAL or NCS list.

Those who follow this sequence avoid the typical late surprises: a wall color not available locally, a faded print, a UI that suddenly looks different in daylight.

What to pay attention to when taking each color sample

Three points are often forgotten in the rush.

First, the screen profile. An uncalibrated laptop may display all tones shifted in one direction. If the palette is shared with printers or painters, your monitor should at least be set to a known profile. An inexpensive probe is sufficient for hobby and studio use.

Secondly, the data source. A palette created from a compressed online image cannot capture all the nuances of the original. If the original is available, it's worth taking the extra step of using it.

Thirdly, human review. Algorithms recognize statistically dominant areas, but not every element of an image. A subtle accent color that carries an image emotionally can be lost in a K-means cluster. The final selection belongs in the hands of the designer, not the software.

Conclusion

A first-class palette doesn't emerge by chance, nor is it created by a single filter. It's the result of a conscious decision regarding the extraction method, the target system, and the subsequent steps in printing, wall paint, or screen. Those who cleanly separate these three levels and work with open, free, and privacy-friendly tools gain reliability for themselves and clarity for clients and workshops. Color thus remains what it should be: a design element, not a gamble.

---

About the author: This article was written in connection with the Shademix project, a free suite of color tools for web, iOS, iPadOS, macOS, Chrome, and Figma. More information, tutorials, and examples can be found at shademix.app.