OpenZhenZhen AI プロンプトギャラリー · gpt-image-2.5 · image

レトロ低画質モノクロカメラで目を隠した自撮り

Retro Low-Res Black-and-White Camera Selfie with Covered Eyes

「レトロ低画質モノクロカメラで目を隠した自撮り」は gpt-image-2.5 を使用した作品です。OpenZhenZhen で元のプロンプトと生成例を確認し、創作の参考にできます。

レトロ低画質モノクロカメラで目を隠した自撮り — AI image creation
レトロ低画質モノクロカメラで目を隠した自撮り

プロンプト

中国語プロンプト

创作一张写实的纵向4:5黑白镜自拍:一位明确成年的东亚女性站在镜子前。将提供的身份参考 ref-10a-v1 作为可见面部身份的唯一依据;在未被遮挡的眼睛和露出的脸颊上保留可辨认特征,不替换或混合身份。深色中长发自然垂落。她将一台浅香槟铜色小型数码相机,机身正面简洁,仅带一处Soran小铭牌举至眼部高度并靠近面部,真实遮住一只眼睛、露出另一只;相机确实位于脸前,不得悬浮。可见眼神放松,略带自我观察感,不强烈凝视。服装:一件宽松炭灰色细格纹针织上衣,带自然衣褶,宽松且不突出身材;胸前配一枚小型深色吊坠。人物置于画面中央偏右。镜面须作为可见成像层,带轻微灰雾、少量细小痕迹和极弱重影。数条柔和浅灰横向反射光带穿过面部与宽松衣物;左侧保留一块较宽的浅灰反光区域,右侧和上方背景压为深炭灰。整体采用克制低反差灰阶、统一柔焦、细腻黑白颗粒、轻微暗角、局部高光泛白及适度细节衰减,营造随手使用旧式小相机自拍的质感,而非锐利商业肖像。将准确文字“Soran”仅印一次,作为相机正面镜头附近的小型深灰实体铭牌;需清楚可读但不抢眼。其他位置不要隐藏文字或微型标记;不要其他可读文字、标志、字幕或水印。保持手部自然、镜面透视可信,并采用安静的卧室式室内环境。

英語プロンプト

Create a photorealistic vertical 4:5 black-and-white mirror-selfie portrait of one clearly adult East Asian woman standing before a mirror. Use the supplied identity reference ref-10a-v1 as the sole authority for her visible facial identity; preserve recognizable features in the uncovered eye and visible cheek without substitution or blending. Her dark medium-length hair falls naturally. She holds a small light champagne-bronze compact digital camera with a plain front and one small Soran nameplate at eye level, close to her face; it physically covers one eye, leaving the other eye visible, and is truly in front of the face rather than floating. Her visible gaze is relaxed and slightly self-observing, not an intense stare. Clothing: a loose charcoal-gray fine-check knit top with natural folds, loose and not figure-emphasizing; add a small dark pendant. Compose the subject center-right. Make the mirror a visible imaging layer with faint haze, a few subtle marks and extremely weak ghosting. Several soft horizontal pale reflection bands cross the face and loose clothing; preserve one broad light-gray reflection area on the left and deep charcoal-gray shadows on the right and upper background. Use a restrained low-contrast grayscale, unified soft focus, gentle black-and-white grain, slight vignette, local highlight bloom and modest detail loss, evoking a casual older compact-camera selfie rather than a sharp commercial portrait. Print the exact text “Soran” once as a small dark-gray physical nameplate on the front of the camera near its lens; it must be legible but unobtrusive. No hidden or microtext marks elsewhere, no other readable text, logos, captions or watermark. Keep natural hands, realistic mirror geometry and a quiet bedroom-like indoor setting.

API モデルを見る

画像・動画プロンプトをもっと見る · 画像を生成 · 動画を生成 · AI キャンバスで作成