Censored Family Snapshot - Image 1

图像提示词

Censored Family Snapshot

Generates a tilted low-resolution candid apartment photo with three people whose faces are obscured by censor blocks, useful for uncanny realistic AI-image demonstrations.

模型
GPT Image 2
分类
Social Media Post
来源
内部整理
9 次浏览 0 次复制 0 次生成点击

完整提示词

可直接复制后粘贴到对应模型,按素材输入情况微调。

返回资料库
Create a low-resolution, candid indoor family snapshot with an unsettling hyper-realistic AI-photo feel, shot at a slight clockwise tilt with a wide-angle phone camera and harsh ceiling-light glare. The scene is a cramped apartment dining/living room with glossy white tile floor, white walls, a bright window with beige curtains in the back left, a dark sofa or cabinet near the window, a wall-mounted water dispenser in the center background, shelves and bottles cluttered in the upper right, and everyday household objects scattered around. Show exactly 3 people: on the left, a seated adult woman in a pale pink long-sleeve top holding a white bowl and chopsticks near the bottom edge; in the center background, a standing child wearing a puffy pink jacket and dark pants, holding something near the mouth; on the right foreground, a leaning adult man in a gray long-sleeve shirt and dark pants, reaching diagonally downward with chopsticks while pointing with his other hand. Cover all 3 faces with opaque rectangular pixelated blur censor blocks in muted brown-gray tones: one tall block over the woman’s face at the far left edge, one smaller block over the child’s face, and one large block over the man’s face. Keep the composition awkward and accidental, with motion blur, compression artifacts, soft focus, noisy shadows, washed-out colors, uneven exposure, and the feeling of an old compressed family photo. Use {argument name="room type" default="small Chinese apartment dining room"}, {argument name="number of people" default="3"}, {argument name="censor style" default="opaque rectangular pixelated blur blocks"}, {argument name="main clothing color" default="pink and gray"}, and {argument name="camera style" default="tilted low-resolution phone snapshot"}. Do not add readable text, logos, watermarks, or extra people.

来源记录

保留原始来源,方便运营复核版权、作者和传播数据。

GitHub 来源:YouMind-OpenLab/gpt-image-2-prompts-search

本页根据 YouMind OpenLab GitHub 开源提示词库重新整理,许可证: MIT。