LabnanaLabnana
Sign In
1 imageGPT-Image-2
以图1作为角色参考生成character a, 以图2作为角色参考生成character b(严格参考妆容、外貌、发色、瞳色、肤色、配色、装饰,角色是动漫角色,不是人偶或真人,体型外貌与年龄毫无关联),2 original characte1 / 21
GPT-Image-2

以图1作为角色参考生成character a, 以图2作为角色参考生成character b(严格参考妆容、外貌、发色、瞳色、肤色、配色、装饰,角色是动漫角色,…

Prompt

以图1作为角色参考生成character a, 以图2作为角色参考生成character b(严格参考妆容、外貌、发色、瞳色、肤色、配色、装饰,角色是动漫角色,不是人偶或真人,体型外貌与年龄毫无关联),2 original character, cute and petite, height difference, character a is taller than character b, 按描述生成图片,內容如下: Anime style illustration, cinematic lighting, romantic atmosphere. Wide shot, significant negative space, blurred and clean urban background. Over-the-shoulder perspective from the right character. character a wearing a sleeveless shirt. She uses her forearms and the backs of her hands to brace against the inside of a dark trench coat, pushing it up to form a protective tent-canopy. She looks slightly surprised and blushing, eyes widened, looking down. Her lower face is partially hidden by character b's head overlapping it, creating strong spatial depth. character b, bare shoulder, seen from behind. Only her cheek profile and ear are visible. She leans in to kiss character a's cheek. Her arm is hanging naturally downwards, elbow relaxed by her side, no arms raised. Her hand is positioned at her own waist level, fingers loosely hooking and holding a single corner of the coat above them, jointly building this shelter. character a leans over character b, showing clear height difference in a protective posture. Soft warm streetlights. Raindrops falling clearly. Highly detailed expressive faces, masterpiece, best quality, 8k. Negative Prompt: messy background, cluttered, extra characters, mutated hands, bad anatomy, distorted faces, extra limbs, missing fingers, blurry, noise, harsh lighting, hands clipping through clothes, hands intersecting fabric, tight hand grips, crossed arms, hands tangled together, arm reaching up, raised arm, arm floating, hands near face, hands in the air, gripping with fingers bent unnaturally, deformed fingers, looking at viewer, eye contact with camera, face fully visible, direct gaze, duplicate characters, characters filling the whole frame.

ModelGPT-Image-2
Resolution2K
Aspect Ratio16:9

Reference Images

Jj z (Hakuno)
4