Replace the **woman in the original video** with the **uploaded person**.
Use the uploaded image only as the **identity reference**. Preserve the uploaded person’s **facial features, gender traits, face shape, eyes, eyebrows, nose, lips, skin tone, hairstyle characteristics, and overall recognizable appearance**. If needed, adapt the body presentation naturally so the final character reads as the uploaded person while staying realistic.
Keep **everything else in the original video the same**, including:
* the original actions and movement
* pose and body motion
* camera angle and framing
* camera movement
* background and environment
* lighting direction and shadows
* video speed and duration
* all objects and scene details
Most importantly, **preserve the exact night-vision look and effect** from the source video:
* green monochrome night-vision filter
* infrared low-light aesthetic
* dark high-contrast atmosphere
* glowing highlights typical of night vision
* visible digital grain / sensor noise
* slight blur or softness from low-light capture
* any vignette, scan, or surveillance-style feeling
* overall eerie night-vision mood
Track the replacement accurately through the whole clip, including head turns, motion, partial side angles, and changing expressions. The uploaded person’s face should blend seamlessly with the original video’s **night-vision texture, lighting, grain, and motion**, without looking pasted on.
Avoid:
* identity drift
* flickering face replacement
* warped facial features
* mismatched skin texture
* loss of the night-vision filter
* over-clean beauty retouching
* changing the environment or camera motion