原始文档
GPT Image 2 · zh
One{argument name="idol theme" default="Dark, yandere, cute idol"} A vertical close-up self-portrait of a girl, set in a narrow, neon-pink room. The camera angle shows her arm outstretched towards the camera, with a shimmering sleeve and a blurred hand visible in the foreground. She wears a richly detailed black and pink plaid stage costume, adorned with glossy vinyl straps, lace trim, ribbons, buckles, and layered accessories: a sleeveless corset, black mesh split sleeves, multiple black armbands on one upper arm, a pink bow, and a heart-shaped badge on her chest inscribed with "病みかわいい" (I feel sick). Her long, dark brown hair is styled in a high bun, with long, straight strands framing her face and decorated with large bows. Her visible hand is close to her lips, striking a shy idol pose, wearing pink lace fingerless gloves with ruffles, long manicures, and chunky rings. A pink armband with handwritten Japanese characters is added to her upper arm, along with delicate earrings. Her face is obscured by a large rectangular mosaic in the center, covering most of her forehead and chin. The background is cluttered and decadent, filled with black and pink graffiti, plastic curtains, plush toys, and distressed decorations. The scene must contain four clearly visible Japanese texts: graffiti on the left reading "闇ー?", below it "嫌いにならないで", and further below that "私だけ見ててね". A glowing heart-shaped neon sign in the upper right corner reads "病みしか勝たん". Two plush toys are placed in the background: a light pink plush animal on the left and a black plush toy with X-shaped eyes on the right. The lighting is a melancholic nightclub-style magenta and violet with glossy highlights, creating a chaotic underground idol dressing room atmosphere. High detail, luxurious texture, glitter, lace, vinyl materials, a messy yet cute gothic atmosphere, high contrast, smartphone selfie composition, and a vertical screen ratio.使用方法
- 复制完整原始提示词，并保留其中的参数。
- 按需求替换主体、文字和构图要求。
- 在兼容的外部工具中使用，核对当前使用条件并检查结果。
开始前需了解
- 先确认是否需要参考图。标注兼容模型，不代表该示例由这个模型生成。
- 核对名称、数字与画面文字；不同模型版本的结果可能不同。
