Menu
All guides|
Color by NumberSep 17, 2026·9 min read

How to Create a Mystery Coloring Book From Photos

A mystery book made from photographs lives or dies at the reduction stage. If the subject survives being flattened into eight colors, everything afterwards is production work.

From Photograph to Mystery Page

PhotographThousands of colors, no regions
Reduced paletteEight colors, flat regions
8
8
8
8
8
8
8
8
8
8
8
8
8
8
5
4
8
8
8
8
8
8
5
5
5
5
4
2
8
8
7
7
5
5
5
4
4
2
7
7
7
7
5
5
4
4
3
2
7
7
7
7
4
4
4
3
2
1
7
7
7
7
2
3
2
2
1
1
7
7
7
7
7
1
1
1
1
7
7
7
7
7
6
6
6
6
6
6
7
7
7
7
7
7
7
7
7
7
7
7
Coded pageWhat the reader is given
AnswerWhat they reveal

The answer image and the reduced-palette image are the same picture. The coded page is that picture with the color taken out and the numbers left behind.

Quick answer

Start from the reduction, not the photograph. The answer image and the reduced-palette image are the same picture, so whether the reveal works is decided the moment the photo is flattened into eight or twelve flat colors — judge every photo by its reduced version, never its original. Choose photos with tonal separation between subject and background and a subject that fills the frame, then pick the smallest palette in which the subject stays recognizable. Merge micro-regions produced by fur, hair and foliage rather than shrinking the codes, fix one palette, color count and minimum region size for the entire book, and collect roughly 140 candidate photos for a 100-page book because photographs fail at conversion rather than at selection. Print and color one page at final size before converting the rest.

Converting one photograph into a mystery page is a solved problem — a tool does it in seconds. Converting a hundred into a book that holds together is not, and the reason is specific to photographs: a mystery book made from drawings starts with artwork that was designed to be colored, while one made from photographs starts with images that were never meant to be reduced at all.

From Photograph to Mystery Page

PhotographThousands of colors, no regions
Reduced paletteEight colors, flat regions
8
8
8
8
8
8
8
8
8
8
8
8
8
8
5
4
8
8
8
8
8
8
5
5
5
5
4
2
8
8
7
7
5
5
5
4
4
2
7
7
7
7
5
5
4
4
3
2
7
7
7
7
4
4
4
3
2
1
7
7
7
7
2
3
2
2
1
1
7
7
7
7
7
1
1
1
1
7
7
7
7
7
6
6
6
6
6
6
7
7
7
7
7
7
7
7
7
7
7
7
Coded pageWhat the reader is given
AnswerWhat they reveal

The answer image and the reduced-palette image are the same picture. The coded page is that picture with the color taken out and the numbers left behind.

A photograph, the same photo reduced to eight colors, the coded page, and the answer

That diagram contains the one idea worth taking from this guide. The answer image and the reduced-palette image are the same picture. The coded page is that picture with the color removed and the codes left behind. So the quality of the reveal is decided entirely at the reduction stage — before any code is assigned, before any page is laid out. Everything downstream can only preserve or degrade it.

If the mystery format itself is new to you, what is mystery coloring covers the mechanics first. This guide assumes you know what a coded page is and concentrates on what photographs add to the problem.

The Reveal Has to Survive Color Reduction

A photograph carries thousands of colors, continuous gradients, and detail at every scale. A mystery page carries eight or twelve flat regions. The conversion throws away almost everything, and the only question that matters is whether the subject is still recognizable in what is left.

This is a stricter test than the one line-art conversion has to pass. An outline conversion keeps edges and discards tone; a mystery page discards edges and keeps tone, in quantized steps. So photographs that convert beautifully into coloring pages can convert into unreadable mystery pages, and the reverse happens too.

The practical consequence: judge every photo by what it looks like after reduction, not before. A tool that shows you the answer image before you commit to the page is showing you the only preview that counts.

Choose Photos for the Reveal, Not for the Outline

Source Photo Quality

Poor Source

Busy background, low contrast

Good Source

Clear subject, strong contrast
A busy, low-contrast source photo beside one with a clear subject and strong separation

The general rules for photo selection still apply — what makes a good photo for color-by-number conversion covers them, and how to prepare a photo for conversion covers the cropping and contrast work beforehand. Two of them matter far more here than elsewhere.

Tonal separation between subject and background. Not color separation — tonal. A red dog against green grass looks well separated to the eye, but if the two are the same brightness they can quantize into adjacent palette entries and the dog dissolves into the lawn. Squint at the photo, or view it in grayscale. If the subject disappears, it will disappear on the page too.

A subject that fills the frame. A small subject in a wide scene has few cells to be built from, and the reveal arrives as a vague smudge. Crop aggressively. It is better to lose context than to lose the subject.

Flat, overcast light is the quiet killer. It is flattering in a photograph and disastrous in a mystery page, because the shading that defines form is exactly what the palette steps are made from. Directional light gives the conversion something to work with.

Palette Size Is Two Decisions at Once

Color count is the main dial you have, and it moves two things in opposite directions.

The Same Photo at Three Palette Sizes

5 colorsEasy to color, least like the photo
9 colorsThe usual working range
15 colorsClosest to the photo, smallest regions

Palette size is one dial controlling two things at once: how much of the photograph survives, and how small the regions get.

The same photograph quantized to five, nine and fifteen colors

More colors means more of the photograph survives — and smaller regions, more of them, and a harder page. Fewer colors means large comfortable regions and a page that drifts further from the original.

PaletteTypical result
5–7Simple, large regions; works for bold subjects and children's books
8–12The usual working range for photo-sourced pages
13–18Detailed, closer to the photo, noticeably slower to color
18+Advanced work, and where near-identical adjacent colors start to appear

There is no correct number, and the question is usually asked backwards. The useful question is not "how many colors should I use" but "what is the smallest palette in which this particular subject is still recognizable?" Start there and add only if the reveal fails.

One constraint worth applying at book scale: every color has to be one a reader can plausibly own. A palette needing six distinguishable greens assumes a pencil set most buyers do not have. How many colors should a color-by-number page have covers the reasoning in more depth.

Micro-Regions Are the Photo-Specific Failure

Illustrations do not produce hundreds of single-cell regions. Photographs do, constantly, and always in the same places: fur, hair, foliage, water, gravel, fabric texture, and dappled light.

These are texture — the photograph's information at a scale the page cannot use. Left alone they produce regions too small to hold their own code, which is the most common reason a photo-sourced page is unpleasant to color. The fix is to merge small regions into their neighbors rather than to shrink the codes, and to accept that texture becomes flat color.

Set a minimum region size as a book-wide rule rather than deciding per page, and set it against the coloring tool you expect readers to use: markers need noticeably larger regions than colored pencils, and fine-tip pens can go smaller but stop being enjoyable long before they stop being possible.

The instinct to fix a muddy page by adding colors or resolution is worth resisting. More detail usually makes a photo-sourced page worse, not better.

Portraits Are the Hard Case

Faces are the most requested subject and the least forgiving. The eye reads a face by relationships measured in fractions of an inch, so a quantization step that shifts a shadow near an eye or the corner of a mouth does not produce a slightly imperfect portrait — it produces a different person.

If your book contains faces, budget review time for them specifically. Inspect the eyes, nose and mouth of every answer image at high zoom before accepting the page, and expect to need a larger palette around the face than the rest of the book would otherwise want. How to turn a portrait photo into a coloring page covers what survives conversion and what does not.

Pets are easier, which is why they dominate this category — a dog is recognizable from silhouette and markings, and small errors in fur read as fur.

One Preset Across the Whole Book

The difference between a converted folder and a book is that a book made a decision once and applied it everywhere.

Fix these before converting page two, and change them only with a reason:

SettingWhy it has to be book-wide
Palette and code systemA shared key can only exist if every page uses it
Color countPage-by-page variation reads as inconsistent difficulty
Minimum region sizeSets how the whole book feels to color
Outline weightThe most visible inconsistency on a shelf of finished pages
Page size and orientationMixed orientations complicate every later layout step

The payoff for a shared palette is concrete: the key gets printed once in the front matter instead of on every page, and a reader working with one set of pencils never has to relearn it. How to create a color by number book for Amazon KDP covers palette locking and answer-key pairing for numbered books in full — all of it applies here.

Plan a Yield, Not a Page Count

Photographs fail at conversion, not at selection, and that timing is what catches people out. A photo that looked fine in the folder reveals its problem only after it has been reduced — by which point you have spent the time.

Plan a Yield, Not a Page Count

140Photos collected
40Rejected
100Pages in the book
Subject too small in frame12
Background too busy to simplify11
Flat light — subject merges with ground9
Reveal unrecognizable after reduction8

Photographs are rejected at conversion, not at selection — which is why the shortfall always arrives late.

140 photographs collected, 40 rejected at conversion, 100 pages in the book

So collect well beyond your target. For a hundred-page book, start from a hundred and forty candidates and expect to lose a quarter to subjects too small in frame, backgrounds that will not simplify, flat light, and reveals that do not survive reduction.

Sorting the survivors by difficulty is worth doing before layout: an easy opening section, a substantial middle, and a harder final stretch gives the book a shape. Grouping by theme does the same job differently. Either beats the order the files happened to be in.

Rights, and What KDP Needs to Know

Two things to settle before publishing rather than after.

Rights to the photographs. Your own photos are usually straightforward; photos taken by someone else are not, and neither are stock images without checking the license for commercial use. Photographs of identifiable people carry privacy and publicity considerations beyond copyright. KDP places responsibility for all of this on the publisher, which means on you.

AI disclosure, if AI is anywhere in the chain. KDP's distinction is between AI-generated content, which must be disclosed, and AI-assisted work, which need not be. Converting your own photograph is a different situation from generating an image from a prompt, but how it should be answered depends on the policy as written when you publish. How to sell coloring books on Amazon KDP covers the disclosure question and account health in more detail.

Convert a single representative page, print it at final trim size, and color it yourself. Not a spot check on screen — the whole page, by hand.

This is the cheapest test in the entire project and the one that catches the most. Screens flatter coded pages badly: codes look legible at 200% zoom and vanish at trim size, regions look colorable that are not, and a palette that looks clearly separated on a backlit display can print as two shades of the same thing.

What you learn from that one page is a set of settings you can then apply to the rest of the book. What you learn from converting a hundred pages first is that you have a hundred pages to redo.

Once the pages exist, the book-level work is the same as for any coloring book: assembling the interior PDF covers trim size, bleed, margins and export, and the coloring book quality control checklist covers the final pass before upload.

Common Mistakes

  • Judging photos before reduction. The answer image is the only preview that matters.
  • Choosing photos for color contrast instead of tonal contrast. Two different colors at the same brightness merge into one region.
  • Fixing a muddy page by adding colors. More detail usually makes a photo-sourced page harder, not clearer.
  • Letting texture become regions. Fur, foliage and fabric produce micro-regions no one can color.
  • Changing settings per page. Converted folders look like folders; books look like decisions.
  • Planning a page count instead of a yield. Photographs fail late, after the work is spent.
  • Accepting portraits without inspecting the face at high zoom. Small errors near the eyes read as a different person.
  • Converting the whole book before printing one page. Everything a print test would have told you now applies to a hundred pages.

Conclusion

A mystery book made from photographs lives or dies at the reduction stage. If the subject survives being flattened into eight or twelve colors, everything afterwards is production work. If it does not, no amount of careful coding, layout or printing will recover it.

So the workflow is: choose a theme, gather far more photos than you need, judge each one by its reduced version rather than its original, fix a palette and a minimum region size once and apply them everywhere, watch the portraits, print one page before converting the rest, and assemble the book the way any coloring book is assembled.

The goal is not a hundred generated pages. It is a hundred pages where the reveal is worth the work — which, for photo-sourced books, usually means fewer pages made from better photographs.

You can turn a photo into a numbered page or into a mosaic coloring page and print one to see how your own photographs reduce before committing to a collection. For personalized books built from pet or family photographs, the custom color-by-number page covers commissioned work, and for a KDP interior the KDP color-by-number generator applies one setting across a whole batch and exports at 8.5 × 11 in and 300 DPI as a PDF, PNGs, or a ZIP.

Printable quality checklist

  • Theme chosen before photos are collected
  • Candidate photos gathered well beyond the target page count
  • Every photo judged on its reduced version, not its original
  • Subject and background separated tonally, not just by hue
  • Subject cropped to fill the frame
  • Smallest palette that keeps the subject recognizable chosen
  • Every palette color one a reader could plausibly own
  • Minimum region size set book-wide, against the expected coloring tool
  • Micro-regions from fur, foliage and texture merged, not shrunk
  • Faces inspected at high zoom in the answer image
  • One palette, code system and outline weight across every page
  • Answer image saved and paired to its puzzle by number
  • Rights confirmed for every source photograph, including people in them
  • AI disclosure question answered against the current KDP policy
  • One page printed and colored at final trim size before converting the rest

Key takeaways

  • The answer image and the reduced-palette image are the same picture — the reveal is decided at reduction.
  • Judge photos by their reduced version, not the original; a good line-art source is not automatically a good mystery source.
  • Tonal separation matters more than color separation — two colors of equal brightness merge into one region.
  • Palette size moves two things at once: how much of the photo survives, and how small the regions get.
  • Fur, hair, foliage and fabric produce micro-regions no one can color; merge them rather than shrinking codes.
  • Portraits are the hard case — a shifted shadow near an eye reads as a different person.
  • Fix palette, color count, minimum region size and outline weight once and apply them across the whole book.
  • Plan a yield, not a page count: photographs fail at conversion, after the work is already spent.

Frequently asked questions

Can I make a mystery coloring book from my own photos?

Yes, provided the photos suit the conversion and you hold the rights to use them. The images that work are the ones whose subject stays recognizable after being reduced to a small palette: a subject filling the frame, tonal separation from the background, and directional light rather than flat overcast light.

How many colors should a mystery coloring book use?

Eight to twelve is the usual working range for photo-sourced pages. Five to seven produces large, comfortable regions but drifts further from the photograph; thirteen or more preserves detail at the cost of smaller regions and a slower page. The useful question is the smallest palette in which your particular subject is still recognizable.

Why does my photo look muddy after conversion?

Usually because the subject and background sit at similar brightness, so they quantize into adjacent palette entries and merge. Check the photo in grayscale before converting. Adding more colors rarely fixes it and usually makes the page harder to color; recropping or choosing a better-lit photo generally does.

Should every page use the same number of colors?

Yes for most books. A shared palette is what lets the key be printed once in the front matter, and it means a reader with one set of pencils never has to relearn the codes. Varying the color count page by page reads as inconsistent difficulty rather than as variety.

How do I stop pages having hundreds of tiny regions?

Set a minimum region size as a book-wide rule and merge anything below it into its neighbors. Micro-regions come from texture — fur, hair, foliage, gravel, fabric — which is detail at a scale the printed page cannot use. Set the threshold against the coloring tool you expect readers to use, since markers need noticeably larger regions than pencils.

Do portraits work as mystery pages?

They are the hardest subject. Faces are read through relationships measured in fractions of an inch, so a quantization step that shifts a shadow near an eye or mouth produces a different person rather than an imperfect likeness. Budget extra review time, inspect faces at high zoom in the answer image, and expect to need a larger palette around them.

How many photos do I need for a 100-page book?

Around 140. Photographs fail at conversion rather than at selection, so the shortfall arrives late — after the work is spent. Expect to lose candidates to subjects too small in frame, backgrounds that will not simplify, flat light, and reveals that do not survive reduction.

Can I sell a mystery coloring book made from my photos?

Yes, when you hold the necessary rights to the source material and the book meets the platform's requirements. Photos taken by someone else, stock images, and photographs of identifiable people all carry obligations beyond your own copyright. KDP places responsibility for all of it on the publisher.

Is a mystery coloring book the same as a color-by-number book?

They overlap heavily and the terms are often used interchangeably. Color by number describes the coding mechanism; mystery describes the withheld subject. A photo-sourced mystery book is usually a color-by-number book whose pages do not show the subject in outline.

MS

Written by

Mosaci Studio Team

The team behind Mosaci Studio

Mosaci Studio is built to help KDP publishers, Etsy sellers, and hobby artists turn photos into print-ready coloring pages without any design experience.

Published Sep 17, 2026

Creating a KDP coloring book?

Batch-convert a folder of images into numbered pages and answer keys, then export a print-ready 8.5 × 11 in interior at 300 DPI.

Or have us produce the book
Explore Mosaci for KDP