AI Redness Analysis — Tone Uniformity, Inflammation, Rosacea Signal
Visible redness from sensitivity, post-inflammatory marks or rosacea is one of the strongest cues that the face does not look 'clear.' LooksMax Scan quantifies it in two color spaces (LAB a*, YCrCb Cr) and maps it across your face.
What gets measured
- LAB a* channel + YCrCb Cr fusion for robust redness detection
- Per-zone tone-uniformity score (cheeks, nose, forehead, chin)
- Heatmap shows diffuse redness and concentrated inflammation
- Baseline normalized to your own skin so different lighting still ranks the right zones
How the AI measures redness
The image is converted to LAB and YCrCb color spaces. The a* channel of LAB encodes the green↔red axis; the Cr channel of YCrCb encodes the chroma red. Both are computed inside the MediaPipe face mask, so background and clothing do not contaminate the signal.
The fused redness map is normalized against your own cheek baseline so a warm-toned face is not penalized for being warm-toned. What you see is the deviation above your face's own median.
FAQ
Is this a rosacea diagnosis?
No. It is a screening tool that surfaces visible redness patterns. Rosacea diagnosis requires a clinician. The scan can be useful to share with one.
Will warm-toned skin always look red?
No. The scan compares against your own facial baseline, not a population mean, so warm-toned skin is not penalized for being warm-toned.
Does makeup affect the score?
Yes. Bare or minimally made-up skin gives the cleanest measurement. The pipeline does not try to undo foundation.
Get your redness map
Upload one photo. The redness pipeline runs in under five seconds and you get a heatmap plus per-zone scores you can compare against future scans.
Start the scan