How the Analysis Works

Powered by open-source research: MediaPipe (Google Research), MODNet (SJTU), and RITnet (MIT).

Your Features

These color samples come directly from your image — the true tones of your skin, hair, eyes, and lips as detected by our analysis.

If a feature looks off — for example, hair appearing darker than expected — try retaking your photo in brighter, neutral lighting.

Undertone Analysis

Our confidence score combines undertone stability (ΔE variation across regions), lighting balance, and cross-feature agreement.

Lower ΔE values and tighter angular alignment between undertone and RAW signals indicate stronger reliability.

How to Get the Best Results

Photo Quality Check

Photography Guidelines

  • Use diffused daylight near a window (no direct sun).
  • Keep background neutral to avoid warm color bounce.
  • Avoid filters and SPF flashback.
  • Keep both sides of your face evenly lit.

Natural, even light and a neutral background give the most stable undertone projection.

Why the Trait-Grid System Matters

Histogram Refinement
Trait Grid Axes
Temp
Depth
Clarity
Human-Consistent Results

Histogram refinement captures complete color distributions — not just averages — while the trait grid interprets temperature, depth, and clarity using multi-region comparison.

Together they produce human-consistent results even under real-world lighting.

Your Season and Traits

Shop Colors