相信大家在使用AI绘画的时候，常常会遇到人物方向总是不对，或者使用抽卡的方式来不断尝试（太费时间了）

那我们应该如何在AI绘画时候，控制人物的视角呢？

其实很简单，我们仅仅需要再Prompt中添加视图关键词，即可让生成的人物（真人、动漫都可以）最终呈现出我们想要的视图

关键词：

front view：前视图

side view：侧视图

back view：后视图

Top view：顶视图

额外分享：Bottom view：仰视图（基本不会出现从鼻孔拍摄的角度，只会让人物更多的抬头），如果大家有尝试出来从下往上拍的视图，可以分享出来哦~

这是我们的基础Prompt：(masterpiece, best quality:1.4), (8k:1.3), ultra-detailed, highly detailed, extremely detailed light and shadow, Tyndall Effect, best light, 1girl, (tohsaka-rin:1.1), with hair dishevelled, (extremely detailed face:1.2), close-up, long hair, babydoll, the study, the bookshelves, the desk full of books, the kerosene lamp, the sun shines into the study, the dust shines in the sun, the leather chair, primitive interior decoration, occult atmosphere, magic and fantasy

这组Prompt中没有提及到任何关键视图的关键词，那么成图时的视图都是随机产生的

只需要我们在Prompt中插入视图关键词：back view（后视图）

(masterpiece, best quality:1.4), (8k:1.3), ultra-detailed, highly detailed, extremely detailed light and shadow, Tyndall Effect, best light, back view, 1girl, (tohsaka-rin:1.1), with hair dishevelled, (extremely detailed face:1.2), close-up, long hair, babydoll, the study, the bookshelves, the desk full of books, the kerosene lamp, the sun shines into the study, the dust shines in the sun, the leather chair, primitive interior decoration, occult atmosphere, magic and fantasy

插入关键词后，我们的成图90%上都是后视图了（成图样例见下方~）

OK，是不是很简单，那么下面就分享一些成图效果把~

前视图样例

Prompt：(masterpiece, best quality:1.4), (8k:1.3), ultra-detailed, highly detailed, extremely detailed light and shadow, Tyndall Effect, best light, front view, 1girl, (tohsaka-rin:1.1), with hair dishevelled, (extremely detailed face:1.2), close-up, long hair, babydoll, the study, the bookshelves, the desk full of books, the kerosene lamp, the sun shines into the study, the dust shines in the sun, the leather chair, primitive interior decoration, occult atmosphere, magic and fantasy,

侧视图样例：

Prompt：(masterpiece, best quality:1.4), (8k:1.3), ultra-detailed, highly detailed, extremely detailed light and shadow, Tyndall Effect, best light, side view, 1girl, (tohsaka-rin:1.1), with hair dishevelled, (extremely detailed face:1.2), close-up, long hair, babydoll, the study, the bookshelves, the desk full of books, the kerosene lamp, the sun shines into the study, the dust shines in the sun, the leather chair, primitive interior decoration, occult atmosphere, magic and fantasy,

顶视图样例

Prompt：(masterpiece, best quality:1.4), (8k:1.3), ultra-detailed, highly detailed, extremely detailed light and shadow, Tyndall Effect, best light, Top view, 1girl, (tohsaka-rin:1.1), with hair dishevelled, (extremely detailed face:1.2), close-up, long hair, babydoll, the study, the bookshelves, the desk full of books, the kerosene lamp, the sun shines into the study, the dust shines in the sun, the leather chair, primitive interior decoration, occult atmosphere, magic and fantasy,

仰视图

Prompt：(masterpiece, best quality:1.4), (8k:1.3), ultra-detailed, highly detailed, extremely detailed light and shadow, Tyndall Effect, best light, Bottom view, 1girl, (tohsaka-rin:1.1), with hair dishevelled, (extremely detailed face:1.2), close-up, long hair, babydoll, the study, the bookshelves, the desk full of books, the kerosene lamp, the sun shines into the study, the dust shines in the sun, the leather chair, primitive interior decoration, occult atmosphere, magic and fantasy,

OK，这次的分享很简单，主需要记住视图的关键词，就可以实现了

样例成图使用的model是：Cetus-Mix_CodaEdition，有兴趣的小伙伴自行下载尝试哦

如果大家不能再本地部署stable diffusion，也可以尝试Google colab云端部署方案

https://huke88.com/article/8051.html