BEFOREAFTERBEFOREAFTER

左右滑動比較

照片旁多一個 3D 迷你分身,一張照片就能試

讓迷你版的你站在身旁,連髮型、穿搭和手上的小物都跟著搭配。用 ChatGPT 上傳照片、貼上提示詞就能試。

這次的照片效果
本人盡量保留同款穿搭與姿勢3D 迷你分身
看實際效果

製作:AI Tip Club · 更新

範例由 AI 生成。結果會因輸入照片、模型與設定而異,提示詞不保證產生完全相同的結果。

怎麼用

  1. 打開 ChatGPT
  2. 上傳一張自己的照片
  3. 貼上英文提示詞,按送出

提示詞(Prompt)

PROMPT · EN
Use the uploaded photo as your only reference. Create a high-resolution image that preserves the real person and the original background as closely as possible, adding exactly one adorable 3D mini-me beside the person.

Preserve the real person's facial features, hairstyle, hair color, expression, body shape, clothing, clothing colors and details, bag, accessories, shoes, handheld items, pose, hand positions, and body orientation. Maintain the original composition, camera angle, lighting, shadows, background, and location. Keep the real person photorealistic without retouching or beautifying them. Do not change the background.

Make the mini-me clearly recognizable as a miniature character version of the same person. Naturally mirror the person's hairstyle, hair color, outfit, colors, accessories, shoes, props, pose, hand gestures, and body orientation. If the person is holding an item, the mini-me must hold a scaled-down version of that same item.

Render the mini-me as a polished, adorable 3D storybook character, not a realistic baby. Preserve the person's likeness while giving the character a cuter, more charming appearance. Include large sparkling eyes, soft cheeks, delicate eyelashes, a small nose, a warm smile, full and richly detailed hair, and dimensional facial features. Use an original character design without copying any specific animated character or recognizable famous style.

Make the mini-me approximately 55–60% of the real person's height. Do not make it too small or place it far away. Position it very close beside the person's legs or body so they look as though they posed for the photo together. Match the ground contact and shadows accurately. Do not obscure the real person's face or important parts of their body.

The final image must keep the real person and the background photorealistic, with only the mini-me rendered as an adorable 3D character that fits naturally into the photograph.

Restrictions:
- Add exactly one mini-me.
- Do not add any other people who are absent from the original photo.
- Do not add pets that are absent from the original photo.
- Do not make the mini-me look like a real baby.
- Do not make it look like a tiny miniature figurine.
- Do not position it far from the real person.
- Avoid awkward or malformed hands, arms, legs, and body proportions.
- Do not add text, watermarks, collages, or split-screen layouts.
- Do not turn the entire image into animation or illustration.

提示詞請直接貼英文原文,效果最好;聊天用中文也沒關係。

再拿 5 種迷你分身變化版

留下 Email,收到這組提示詞的 5 種變化版,以及每週精選提示詞。下次想玩新效果,直接打開信箱找靈感。

實際效果

長按照片可以看原圖滑鼠移到照片上可以看原圖

AFTER: 用 ChatGPT 在照片旁加一個 3D 迷你分身BEFOREAFTER
AFTER: 用 ChatGPT 在照片旁加一個 3D 迷你分身BEFOREAFTER
AFTER: 用 ChatGPT 在照片旁加一個 3D 迷你分身BEFOREAFTER
AFTER: 用 ChatGPT 在照片旁加一個 3D 迷你分身BEFOREAFTER
下一個提示詞疲憊英雄電影感人像|一張照片就能做看下一個提示詞 →

效果更好的小技巧

  • 選臉、髮型和穿搭都清楚的照片;想讓分身穿同款鞋子,原圖也要看得到鞋子。
  • 優先用單人全身照或半身照,身旁留點空間,方便放入分身。
  • 提示詞請直接貼英文原文,聊天用中文也沒關係。
  • 不同模型、訂閱方案與原圖會影響結果;生成後檢查本人的臉、背景,以及分身的手和小物。

你可能也會喜歡

下一篇 →

先存起來,之後再用

傳給自己,要用的時候馬上找得到。