图像编辑任务。以图 1 作为基础图像。重现相同的场景、姿势、相机角度、构图、光照、服装风格和背景。仅将人物身份更改为与附加的 AI 模…

exact prompt I use to recreate images with ChatGPT 2 image: "Image editing task. Use Image 1 as the base image. Reproduce the same scene, pose, camera angle, framing, lighting, outfit style and background. Change only the person identity to match the attached AI model reference images. Keep the same body orientation, hand position, facial direction and composition. Preserve the raw phone-photo look: slight motion blur, JPEG compression, sensor noise, imperfect crop, real shadows, uneven indoor light and natural skin texture. Do not make it look like a studio render." The goal is not a new image. The goal is the same image with my model inside it.

图像编辑任务。以图 1 作为基础图像。重现相同的场景、姿势、相机角度、构图、光照、服装风格和背景。仅将人物身份更改为与附加的 AI 模特参考图匹配。保持相同的身体朝向、手部位置、面部方向和构图。保留原始手机拍照的质感：轻微的运动模糊、JPEG 压缩、传感器噪点、不完美的裁剪、真实的阴影、不均匀的室内光线以及自然的皮肤纹理。不要让它看起来像工作室渲染图。 目标不是生成一张新图像。 目标是同一张图像中放入我的模特。