Gpt Image 2 On Chatgpt Prompt; Use Image

the prompt

Use image 1 as the sole identity reference and preserve the subject’s facial features, proportions, skin tone, and overall appearance exactly. Create an ultra-photorealistic 4:5 mirror selfie inside a refined luxury dressing room. The woman is captured in a deep, natural squat in front of a tall full-length mirror with a warm wooden frame. She holds a light-colored smartphone in her right hand, partially obscuring her face, while her left hand carries a dark brown leather clutch. She is dressed in a flowing sand-beige open-weave crochet maxi dress with delicate shoulder straps and a plunging V-shaped neckline, without any waist belt. Complete the outfit with cream-colored stiletto sandals, oversized statement bracelets, elegant drop earrings, and sleek, straight hair falling naturally over her shoulders. The room features soft neutral walls, a gray carpet, and diffused natural daylight that creates realistic HDR lighting, true-to-life reflections, lifelike skin texture, accurate anatomy, and the feel of a candid high-end fashion photograph.

Negative prompt: different face or identity, facial distortion, plastic-looking skin, unrealistic anatomy, malformed or extra fingers, fused limbs, incorrect mirror reflection, warped proportions, belt added, incorrect outfit, missing clutch, low quality, cartoon, CGI, text, logo, watermark.