Methodology

How AscendMe actually scores your face.

This page documents exactly how AscendMe turns two photos into a bounded, cited, non-surgical aesthetic report.

1. Landmarking

140+ 2D facial landmarks are extracted from both the front and side photo. Landmark positions drive every downstream measurement.

2. Structural measurements

From landmarks we compute canthal tilt, facial thirds, fifths, midface ratio, gonial angle, lip-to-chin, philtrum length, cheekbone projection, jaw width, and bilateral asymmetry.

3. Skin & health signals

Skin clarity, texture, tone uniformity, under-eye contrast, lip hydration, and eye brightness are scored from the front photo.

4. Calibration

Every metric is compared to a demographic-appropriate reference distribution using your reported age, gender, and regional background.

5. Human QA

A trained reviewer verifies the AI's outputs, flags anything unrealistic, and confirms the roadmap makes sense for what the photos show.

6. Reporting

The final report presents a composite 0–100 score, six sub-scores, a bounded before-vs-achievable visualization, top three priorities, and a phased 0–3 / 3–6 / 6–12 / long-term plan.

Common questions

What research underpins the scoring?

Peer-reviewed craniofacial and aesthetic research on facial thirds and fifths, canthal tilt, gonial angle, midface projection, dimorphism, and skin-quality metrics.

Is the model calibrated per demographic?

Yes. Age, gender, and regional background all adjust the reference distributions.

Does a human review the report?

Yes. A trained reviewer signs off on every report before it reaches you.

Ready when you are.

Upload two photos. Get your bounded, science-backed roadmap in 24–48 hours.

Start my analysis

Explore more