Photorealistic soft-girl selfie portrait with face reference in GPT Image 2

Use the uploaded reference image as the primary facial reference. Preserve the exact facial identity, facial structure, facial proportions, eye shape, nose shape, lip shape, jawline, cheekbones, skin tone, natural asymmetry, hairstyle details, and overall appearance with maximum accuracy. Do not alter age, ethnicity, or any recognizable facial characteristics. A close-up portrait of the same young woman from the reference image with a fit, athletic appearance. Hair styled dark brown and secured with a clip, with long soft bangs framing both sides of her face. Distinctive soft-girl makeup featuring heavy rosy-pink blush across the cheeks and nose bridge, naturally plump matte pink-nude lips, subtle freckles scattered across the nose, long natural eyelashes, and wide expressive eyes looking directly into the camera. Natural skin texture with visible pores and realistic detail. She is wearing a light sky-blue ribbed cotton tank top with thin spaghetti straps and a casual fit. Accessories include thick gold statement hoop earrings and a delicate gold snake-chain necklace resting on the collarbone. Head-and-shoulders composition, facing forward in a confident, slightly sassy pose, head gently tilted to the right. Captured from a slightly low upward angle to enhance the selfie perspective. Bright daytime indoor setting with clean aesthetic white walls in the background. High-quality smartphone selfie aesthetic, flash photography, social-media video screenshot style, sharp focus on the eyes and lips, subtle cinematic soft focus, ultra-realistic skin rendering, natural lighting, shallow depth of field, photorealistic, ultra-detailed, 8K quality. Negative Prompt: (text, UI elements, watermark, social media overlay:1.5), (heavy contouring:1.3), sharp artificial makeup, altered facial identity, different face, different hairstyle, unrealistic skin smoothing, plastic skin, doll-like appearance, distorted anatomy, bad teeth, messy fingers, extra limbs, cartoon, illustration, 3D render, anime, painting, overly dramatic lighting, excessive saturation, grain, blur, low resolution, closed eyes, looking away from camera.