Keep the uploaded person’s face EXACTLY unchanged. Preserve identity 1:1.

IMPORTANT:

Do NOT remove, replace, or alter any surrounding subjects (people, dogs, objects).
All nearby individuals and animals must remain present in the exact same positions.
Maintain the full composition — no cropping, no simplification of the scene.
Do NOT reduce the number of subjects in the image.

Scene:
A real KBO live broadcast (SPOTV style), shot from stadium spectator seats, medium zoom.

Environment:
Crowded stands with multiple spectators, visible people sitting nearby, dogs sitting on adjacent seats, beer cups, cheering props, handheld fans. Busy and slightly messy stadium atmosphere.

Pose & behavior:
The person is sitting casually, watching the game, unaware of the camera. Neutral candid expression.
Dogs and nearby people are naturally seated and behaving casually (no posing, no staging).

Visual style (IMPORTANT):
Real broadcast capture look:

slight motion blur
compression artifacts
mild noise
imperfect lighting
natural skin texture (no smoothing)

NEGATIVE constraints:

do NOT remove dogs
do NOT remove surrounding people
do NOT isolate the main subject
no beauty enhancement
no face modification
no cinematic or stylized look
no AI-clean composition

Keep the uploaded person’s face EXACTLY unchanged. Preserve identity 1:1.

Scene:
An ordinary spectator accidentally captured on a real KBO live broadcast (SPOTV style).
Shot from stadium spectator seats, medium zoom.

Environment:
Crowded stands, people around, beer cups, cheering props, handheld fan visible.
Natural, slightly messy stadium atmosphere.

Pose & behavior:
Sitting comfortably with legs crossed, watching the game.
Unaware of the camera. Neutral, candid expression.

Visual style (IMPORTANT):
Real broadcast capture look:
- slight motion blur
- compression artifacts
- mild noise
- imperfect lighting
- natural skin texture (no smoothing)

NEGATIVE constraints:
- no beauty enhancement
- no eye enlargement
- no jawline correction
- no skin glow
- no makeup enhancement
- no influencer / photoshoot look
- no cinematic grading

Goal:
Must look like a random real person caught on live TV,
NOT like an AI-generated or styled image.