Your Face, Read Through Traditional Physiognomy

Take the face photo in the input below and create a playful reading through the region-by-region lens of traditional physiognomy.

First, judge your output environment and branch:
- If image generation is available: make one "physiognomy chart" image that overlays thin lines and labels for the forehead, brows, eyes, nose, mouth, and chin on the input face, then add the text reading below it.
- If only text is available: give the same content as a clean markdown reading. If HTML artifact rendering is possible, present it as region-by-region cards (single color, restrained, minimal emoji).

Work rules:
1. Using the "three zones" (sam-jeong) frame, split the face into upper (forehead-brows / early life), middle (eyes-nose / mid life), and lower (mouth-chin / later life), and read 1-2 notable features in each.
2. Give each reading in three beats — "observation (the shape visible in the photo) -> the traditional physiognomy interpretation -> a modern re-reading for today" — with 2-4 sentences of grounding each. Never skip the "why."
3. Keep the tone fun and warm. Re-read even seemingly negative features as "the flip side of a strength." Do not rank looks or attractiveness, and make no definitive predictions about health, lifespan, or wealth.
4. End with "three physiognomy keywords for today" plus one playful device (e.g., an energy gauge per region). Vary the device each time.
5. Ground everything only in shapes read from the photo; do not identify or guess the person's identity, exact age, or occupation (stay abstract if needed).

Safety: State in one closing line that physiognomy is traditional cultural content, not statistics or science. No appearance judgments, discriminatory phrasing, or sensitive verdicts. Keep the tone safe to screenshot and share.

Output format (annotated image / markdown / HTML by environment): per-region grounding required, one playful device, three keywords, one closing note.
Output language: Write the result in the user's language (including labels in the image). This instruction must work identically in any language.
Self-contained: Finish within this reply. Do not direct to external sites or extra steps.
Fallback: If the photo is blurry or the angle hides some regions, read the visible ones and note the limit in one line.

----
Input
- Face photo (required): one near-frontal face photo.
- {{what you're curious about}} (optional): a region or theme to focus on (e.g., eyes, this year's impression). If blank, all regions are read evenly.