About this skill

What it solves

Many "attractiveness analysis" tools return a single opaque score. This skill turns a frontal face photo into inspectable dimensions: three-court and five-eye proportions, facial symmetry, golden ratio, skin quality, and a weighted overall score. It is useful when subjective appearance notes need to become structured results.

How it works

The analysis uses MediaPipe Face Mesh and 468 facial landmarks. The workflow is: load a clear, frontal face image, detect landmarks, perform geometric measurements, and use statistical deviation as part of scoring. The golden ratio check compares facial segments against 1.618; symmetry compares left/right differences; skin quality evaluates brightness uniformity and texture fineness. Output includes two parts: a readable formatted report and structured JSON data between [JSON_OUTPUT_BEGIN] and [JSON_OUTPUT_END] markers.

Boundaries

Use frontal, clear, evenly lit photos and avoid glasses or masks, which can reduce accuracy. This is intended as a technical experiment or entertainment reference, not a medical diagnosis or professional aesthetic certification. For batch analysis, standardize input conditions first, then compare downstream using the JSON fields.

Screenshots

Use Cases

Score front-facing model photos in batch and compare golden ratio and symmetry differences
Cast photo review with three-court five-eye, symmetry scores, and JSON details
Check frontal photo clarity and skin quality across a local image folder
Convert playful test results into structured JSON for downstream comparison

Best For

Vision engineer: tuning proportion scoring after Face Mesh landmark detection
Face-data product manager: defining photo scoring dimensions and JSON field specs
Content platform operator: screening model front photos by symmetry and skin quality
Automation engineer: parsing JSON_OUTPUT markers to render report pages