Methodology
Methodology and limitations
MBTI Face Analyzer compares measurable facial-landmark ratios from live camera frames or selected photos with entertainment-oriented reference vectors. It is not an MBTI test, personality assessment, psychological assessment, medical tool, or appearance rating system.
1. Face landmark detection
The browser detects facial landmarks from live camera frames or a selected photo and calculates relative positions of the eyes, nose, mouth, jawline, and face outline. Camera video and photo files are not uploaded.
2. Ratio normalization
To reduce the effect of image size and resolution, the tool converts measurements into ratios based on face width and height. Live capture averages feature values from multiple good frames.
3. Reference vector comparison
The extracted feature vector is compared with 16 MBTI-style vibe reference vectors to find the closest candidates.
4. Safe interpretation
The result is limited to entertainment use and is grounded in extracted features and score differences, not real personality claims.
Main extracted features
Face height-to-width ratio
A normalized value used to describe the overall face outline in the result explanation.
Eye angle
A calculation based on relative eye corner positions.
Eye openness ratio
Eye height compared with eye width, used as one visual cue.
Mouth corner line
The relative position of mouth corners and center line, used as a light expression cue.
Jaw width ratio
Jaw width compared with face width, used to describe facial-line stability.
Cheek-to-jaw ratio
The difference between wider face outline and jawline width, used as a contour cue.
Image quality check
If a face is strongly rotated or tilted, landmark ratios can be distorted. Live capture guides face position and tilt, then averages good frames. Uploaded photos use symmetry and tilt estimates to show a quality warning.
AI explanation scope
The AI does not inspect the photo. It receives only local analysis results such as top candidates, scores, and extracted feature explanations. It is instructed not to invent unsupported claims about personality, intelligence, fortune, gaze, or mental state.
Limitations
Results can vary depending on lighting, expression, lens distortion, hair, masks, filters, and compression. The result should be used only as a light conversation topic.