Use the user’s uploaded reference image as the sole and absolute source of identity for the person being generated.
Preserve exact facial structure, bone proportions, age, ethnicity, asymmetry, and natural features from the reference image.
Do not beautify, idealize, stylize, or reinterpret the face in any way.
No changes to gender, age, skin tone, facial anatomy, or expression beyond what naturally results from lighting.
The face must remain fully recognizable and consistent with the reference image under all conditions.
2. Cinematic Style Transfer
Apply the visual style of the analyzed image: a dark, minimalist, psychological cinematic aesthetic with a strong sense of anonymity and tension.
Overall mood is somber, oppressive, and introspective, evoking isolation within a crowd.
Color palette is extremely restrained, dominated by deep blacks, charcoal gray, muted slate tones, and subtle desaturated skin tones.
Lighting is low-key and directional, with a soft but controlled frontal light hitting only the central face, while surrounding figures fall into near-total shadow.
Contrast is high but not harsh—shadows are deep and clean, highlights are soft and subdued, never glossy.
No dramatic color grading—keep tones natural, cold-neutral, and cinematic, avoiding warm highlights or stylized tints.
3. Composition & Framing
Medium-tight portrait framing, chest-up to shoulders, centered composition.
The subject is positioned perfectly in the center, surrounded closely by multiple human silhouettes wearing dark hooded garments.
Surrounding figures form a dense, repeating pattern, slightly staggered in depth, creating visual rhythm and claustrophobia.
The background is not a location but a mass of human forms, fading into darkness with no identifiable environment.
Clear subject separation achieved through light placement, not background blur or color contrast.
4. Texture & Detail Control
Skin texture must be highly realistic, showing pores, fine lines, subtle under-eye darkness, and natural imperfections.
No AI smoothing, plastic skin, beauty retouching, or artificial sharpening.
Clothing texture (hooded fabric) should appear matte, heavy, and soft, absorbing light rather than reflecting it.
Faces of surrounding figures remain intentionally obscured, with minimal readable detail—only shape and form.
5. Optical & Camera Feel
Cinematic full-frame lens feel, approximately 50–85mm equivalent, neutral perspective with no distortion.
Shallow-to-moderate depth of field: the central face is in crisp focus, while surrounding figures fall slightly out of focus with gentle falloff.
Focus roll-off is smooth and natural, not abrupt.
Overall sharpness is controlled and filmic, avoiding digital over-clarity.
No lens flares, no stylized bloom, no exaggerated bokeh.
6. Strict Constraints
No fantasy elements.
No surrealism or symbolism beyond what naturally emerges from composition and lighting.
No additional props, no environmental storytelling beyond the crowd itself.
No changes to clothing style unless directly implied by the reference image.
No exaggeration of emotion—the expression remains restrained, neutral, and intense.
The final image must feel grounded, realistic, and psychologically cinematic, not illustrative or conceptual.