posts: 添加Gemini Nano Banana AI生图教程及示例图片

添加了关于Google AI Studio生成图片的教程文章，包含立绘转真人、七夕生草图、去除背景等示例

src/content/posts/gemini-nano-banana.md

@@ -0,0 +1,86 @@
+---
+title: Google最新的AI生图模型！立绘转真人！一键P！
+published: 2025-09-01T16:25:34
+description: ''
+image: ''
+tags: []
+
+draft: false 
+lang: ''
+---
+
+# 快速上手
+
+进入 [Google AI Studio](https://aistudio.google.com/prompts/new_chat)
+
+粘贴一张图，说出你想要的更改
+
+![](../assets/images/2025-09-01-16-33-14-1dc081b4dcb4eb99fd68c1d97322517a.png)
+
+# 立绘转真人COS
+
+> Prompt：Generate a highly detailed photo of
+> 
+> a girl cosplaying this illustration, at Comiket.
+> 
+> Exactly replicate the same pose, body posture,
+> 
+> hand gestures, facial expression, and camera
+> 
+> framing as in the original llustration. Keep the
+> 
+> same angle, perspective, and composition,
+> 
+> without any deviation 
+
+注意：
+
+- 立绘最好为纯色背景，人物主体完整
+
+- 立绘中角色不要做出现实中无法做出的动作，这可能会触发AI的Copy，会将图片完整的复制过来而不是创建一个Coser
+
+![](../assets/images/2025-09-01-16-34-07-7d28242d1b8c18abf6d6b44f8064abff.png)
+
+![](../assets/images/2025-09-01-16-37-48-572b1752-d279-4fc3-93a4-165a75640d84.png)
+
+# 七夕生草图
+
+> Prompt： A photorealistic scene inside a McDonald’s restaurant.
+> 
+> On my table, there is a picture frame showing the anime-style image I provided.
+> 
+> Everything outside the frame must be 100% realistic and human.
+> 
+> On the opposite table, a real Chinese couple is sitting together.
+> 
+> The female partner is cosplaying the character from the provided image, wearing the same outfit, but she must look like a real human girl, not a drawing or cartoon.
+> 
+> ⚠️ She should not mimic the exact pose or gesture from the anime illustration — instead, she should be doing something natural for a restaurant setting, such as eating food, drinking from a cup, talking with her partner, or smiling casually.
+> 
+> The entire environment (restaurant interior, people, cosplay) must look like a natural camera photo.
+> 
+> Apply background blur (bokeh effect) to the opposite table and everything behind it, making the image look like it was taken with a DSLR camera, not generated.
+> 
+> Only the picture inside the frame can contain anime style; all other elements must be realistic.
+
+![](../assets/images/2025-09-01-16-38-35-cfdbd266c99acc5da9bc6d5f4edf57ba.png)
+
+![](../assets/images/2025-09-01-16-39-17-5976a1f7-3f14-4987-aa60-3257b05f8b7b.png)
+
+# 去除背景
+
+> Prompt： 将这张图去除背景，仅保留人物主体
+
+![](../assets/images/2025-09-01-16-39-56-1dc081b4dcb4eb99fd68c1d97322517a.png)
+
+![](../assets/images/2025-09-01-16-40-02-72ff1a60-c98b-443e-b231-e068a0429fb8.png)
+
+---
+
+# 结语
+
+- 最好使用英语Prompt
+
+- 可以先对通用大模型（如ChatGPT）描述你要绘制什么图片，然后让他给你返回英文Prompt
+
+- 好图秀图不是一蹴而就的，需要细心打磨