COLOR LAB / LOCAL TOOL
Choose one JPEG, PNG, or WebP image to begin.
PRACTICAL GUIDE
Look for information that becomes hard to separate. Do not judge the design only by whether the colors look different.
01 / SUPPORT AND SCOPE
It is a model-based design preview. It is not a diagnosis or an accessibility pass/fail test.
02 / HOW TO USE IT
Upload a representative screenshot, chart, map, illustration, or photo.
Select the vision types that matter to your review.
Choose a method and move severity through the range.
Read Original, Simulated, and Difference together.
Check whether labels, symbols, lines, states, and regions remain distinguishable.
METHOD GUIDE
Keep the image, vision type, and severity unchanged when comparing methods.
Start here for a full-dichromacy reference or a tritan comparison. This implementation uses two projection planes and blends toward the result below severity 1.
Use it as a simpler protan or deutan cross-check. Tritan remains selectable, but this tool marks it as a warning case.
Use it when severity is part of the question. This implementation uses severity-specific matrices. Its tritan result is an approximation.
Use it when you need the DeficiencyView label. In this tool it uses the same Machado transform, so agreement between these two options is not independent confirmation.
Achromatopsia note: this view uses luminance grayscale and does not depend on the selected simulation method.
READ THE OUTPUT
Use it to locate the label, state, line, or region you are reviewing.
Look for elements that merge, fade, or lose a color-only distinction.
Brighter areas changed more. This grayscale map lets you inspect change without relying on the same colors being simulated.
Read each map on its own: Difference is auto-scaled separately for every result. Use it to compare the location and shape of change inside one card, not absolute brightness between cards.
03 / INTERPRETATION
Ask whether a chart series, map route, status, warning, or control still has a non-color cue. Labels, position, shape, pattern, and line style can preserve meaning when colors converge.
If two methods disagree, treat that disagreement as uncertainty. Inspect the design and strengthen its non-color cues instead of choosing the most dramatic output.
04 / LIMITS
Color perception varies between people. The screen, ambient light, and viewing conditions also affect what is seen.
Thin marks and large uniform regions may not behave the same way. Test the full design and a closer crop when both carry meaning.
Use a separate contrast checker for text and interface contrast. Test with people when the decision has real consequences.
COMMON QUESTIONS
No. The browser decodes the image, draws it to a local canvas, runs the simulation, and creates the PNG download on this device.
JPEG, PNG, and WebP are accepted. This version processes one image at a time.
Zero keeps the original colors. One applies the full selected model. Values between them show intermediate output. The number is a review setting, not a personal diagnosis or severity score.
If you have a color vision deficiency, both images still pass through your own vision. A color-only comparison may hide the change. Read the Difference map, which expresses pixel change as brightness.
No. Each map is normalized to its own largest pixel change. Compare where change appears inside a map, not how bright one card looks beside another.
They use the same Machado transform in this tool. DeficiencyView is a compatibility preset here, not a second independent model.
It can reveal likely color collisions. It does not issue a pass or fail. Add labels or other non-color cues, check text contrast separately, and include user testing when the stakes justify it.
Thin lines, small symbols, and large uniform areas can reveal different problems. Review a representative full design, then crop the details that carry meaning.
No. This is a design preview, not a medical test. A qualified eye-care professional can assess color vision.
Brighter pixels changed more. The map is auto-scaled for this view.