The result, explained.
AES is Onyxa Medical’s proprietary outcome-documentation metric for structured before-and-after review. It brings AI image grading and available human assessments into one readable case summary without replacing clinical judgment.
One case summary. Six distinct measures.
The gallery intentionally shows several measurements together. They are related, but they are not interchangeable. AES, overall improvement, feature grades, GAIS, and perceived age are separate outputs. Human ratings are incorporated only when collected. AES supports documentation and case review; it does not determine diagnosis, candidacy, or treatment.
How the analysis works.
AES compares the submitted image pair within selected treatment zones. The active production workflow is AI-based; older descriptions of a Gabor-filter or landmark-only scoring path do not describe the current engine.
The rubric follows the zone.
A single generic feature list is not applied to every photograph. Face and body cases use different descriptors while retaining the same lower-is-less-severe 0–4 grade direction. Face analysis uses wrinkles, volume, laxity, and skin quality. Body analysis uses zone-specific descriptors for skin laxity, texture or zone attribute, contour definition, and surface quality.
Two calculations, kept separate.
The overall percentage describes grade change. The AES score maps that percentage to a 0–10 scale and then conditionally incorporates available human ratings. A one-grade reduction equals 25% feature improvement. The overall percentage is an adaptive weighted combination, not a fixed average of every pillar. The current AI score mapping is nonlinear and capped at 10. It is a proprietary scoring transform, not another percentage.
A separate 0–4 severity measure.
Feature grades and AES scores answer different questions. AES summarizes the overall outcome on a 0–10 scale; feature grades describe the visible severity of individual analyzed features before and after treatment. When a case includes grade data, the values remain on the 0–4 severity direction. A lower after value indicates less visible severity for that analyzed feature.
How GAIS enters the score.
In the current workflow, the physician rates the case before the AI score is revealed. The patient rating is optional. Both use the same 1–7 numeric direction and are normalized to 0–10 before any weighting. Physician GAIS is a blinded clinical assessment included only when collected. Patient GAIS is an optional patient-reported perspective included only when collected.
A summary band, not a diagnosis.
The band is a plain-language label for the final AES value. Read it alongside the source photographs, overall percentage, feature changes, GAIS availability, treatment details, and clinician assessment. Do not compare scores across cases as if every photograph, treatment zone, treatment target, timing interval, or GAIS input were identical.
What AES can—and cannot—tell you.
AES is appropriate for structured review of standardized before-and-after photographs, documenting visible changes within selected treatment zones, combining available image and human assessments into one case summary, and supporting clinician-patient discussion when used with source images and clinical context. It is not a substitute for clinical examination, diagnosis, candidacy assessment, treatment planning, standardized photography, independent clinical validation of the complete proprietary AES composite, or a guarantee of outcome, durability, safety, or patient satisfaction.