Midjourney连续人物和角色生成方法: 3种全新的Prompt技巧

小薇 Official Channel
30 Aug 202306:03

TLDR大家好，欢迎来到小薇的频道。在本期视频中，小薇介绍了一种新的prompt技巧，用于Midjourney图像生成软件，以生成连续性的人物图像。通过添加特定的提示词，如'image split into 2'，可以将图像分割成多个区域，每个区域中的人物和背景保持一致，从而实现连续性图像的生成。此外，小薇还展示了如何通过添加描述人物表情、服装、背景的词条来提升生成效果。视频中还推荐了SegMind和Tensor.art作为Midjourney的替代绘图平台，并提供了如何使用这些平台的指导。最后，小薇还介绍了如何使用Midjourney设计矢量图插画，并使用inpainting工具替换人物服装。视频结尾，小薇鼓励观众点赞、订阅并转发频道，并邀请大家留言或联系她。

Takeaways

  • 🎨 Midjourney 可以通过文本生成高质量图像，但生成连续角色图像存在挑战。
  • 🔍 用户Julie W Design发现通过添加特定的prompt技巧，可以基于同一角色生成不同角度和背景的图像。
  • 📐 使用'split'词条可以把一张图像分割为多个区域，实现连续性图像生成。
  • 🐕 可以设计不同动作的柯基犬图像或人物图像，保持高度一致性。
  • 🌟 在提示词中添加描述人物表情、服装、背景的词条，可以生成更出色的效果。
  • 🎭 该方法同样适用于动漫人物的生成，调整模式为niji5，设置分割区域为九个。
  • 🆓 对于没有Midjourney付费账号的用户，推荐使用SegMind和Tensor.art这两个免费的绘图平台。
  • 📈 SegMind提供多种Stable Diffusion模型，用户可以输入提示词并添加特定的词条生成图像。
  • 🖼️ Tensor.art提供丰富的模型和足够的点数，使用户可以方便地生成图像。
  • 📝 使用AnimateDiff工具可以把不同姿势的图像合并成动画。
  • 🎭 Midjourney也可用于设计矢量图插画，广泛应用于出版、媒体、广告等领域。
  • 🔄 通过改变提示词中的变量，可以在图像中添加不同的物体或改变设计风格。
  • 🛍️ 设计好的矢量图可以在WireStock、AdobeStock等平台出售。
  • 👗 Midjourney的新工具inpainting可以用于替换人物服装风格。

Q & A

  • Midjourney 是如何使用文本生成高质量图像的？

    -Midjourney 使用了扩散模型来完成图像的生成，通过文本提示（prompt）来引导图像的生成过程，从而创造出高质量的图像。

  • 为什么 Midjourney 在生成连续角色图像方面存在困难？

    -Midjourney 由于缺乏 inpainting 和 controlnet 插件，因此难以完成连续角色的生成，因为这些插件可以帮助在图像生成过程中保持角色的一致性。

  • Julie W Design 发现的 prompt 技巧是什么？

    -Julie W Design 发现，在提示词中添加 'image split into 2' 可以把一张完整的图像分割为两个区域，每个区域中人物的面容和背景都会保持一致，从而实现连续性图像生成。

  • 如何使用 split 词条来生成连续性图像？

    -通过在提示词中使用 split 词条，可以将同一张图像划分为多个区域，同时设置拍摄角度和分割区域的比例，实现真正意义上的连续性图像生成。

  • 生成连续人物图像时，如何调整表情和背景？

    -在提示词中添加描述人物表情、服装、背景的词条，可以生成具有特定表情和背景的连续性图像。例如，将分割区域调整为六个，添加描述人物表情的词条，就可以生成一组连续性的画面。

  • 动漫人物的连续性图像生成如何操作？

    -将模式调整为 niji5，输入同样的提示词，并将分割区域设置为九个，可以生成动漫人物的连续性图像。

  • 如果没有 Midjourney 付费账号，有什么替代方案？

    -对于没有 Midjourney 付费账号的用户，可以尝试使用新的绘图平台 Segmind，它提供了多种 stable diffusion 模型，并且新用户会获得100点免费额度。

  • 如何使用 Segmind 平台生成图像？

    -在 Segmind 平台上，点击右上角的 'login' 按钮使用 Google 账号登录，然后点击 'try model' 直接使用多种 stable diffusion 模型。输入提示词，并在前方添加特定的词条，选择风格和设置参数后点击生成按钮即可。

  • 除了 Midjourney，还有哪些平台可以生成图像？

    -除了 Midjourney，还可以使用 Tensor.art 平台生成图像，它提供了足够的点数和丰富的模型，使用起来更加方便。

  • 如何使用 AnimateDiff 工具创建动画？

    -使用 AnimateDiff 工具可以将不同姿势的图像合并成动画，用户可以根据自己的需求进行尝试和操作。

  • Midjourney 如何用于设计矢量图插画？

    -Midjourney 可以设计矢量图插画，通过输入特定的提示词公式，其中包含描述图像主要内容和背景的词条，可以生成具有特定风格和元素的矢量图插画。

  • 如何使用 Midjourney 的 inpainting 工具替换人物服装？

    -首先输入一段提示词生成人物图像，然后点击图像下方的 'vary region' 选中人物身体区域，更换原有的服装风格词条，就可以为人物添加任意风格的服装。

Outlines

00:00

🎨 Image Generation Techniques with Midjourney and Stable Diffusion

The video introduces viewers to the channel and discusses the limitations of Midjourney and stable diffusion models in generating continuous character images due to the absence of inpainting and controlnet plugins. However, a prompt technique discovered by a user named Julie allows for the generation of images of the same character from different angles and backgrounds. The host, Xiaowei, demonstrates how to use this technique by showing how to split an image into multiple regions with consistent facial features and backgrounds. The method can be applied to create a sequence of images with varying actions or expressions. The video also covers the use of additional descriptive terms in the prompt to enhance the image generation process. Furthermore, the host recommends a new drawing platform, Segmind, which offers a free trial and access to various stable diffusion models. The video concludes with a guide on how to use Midjourney for vector illustrations, which are widely used in publishing, media, and advertising, and provides a simple formula for creating vector-style icons.

05:01

🖼️ Selling AI-Generated Vector Art and Using Midjourney's Inpainting Tool

This paragraph discusses the commercial aspect of AI-generated art, mentioning that platforms like WireStock and AdobeStock support uploads of AI-created images. The host references a previous video that explains how to use WireStock and encourages viewers to watch it via a provided link. The video also highlights Midjourney's new inpainting tool, which can be used to change a character's clothing style in a generated image. The host demonstrates how to use the tool by selecting a body region and altering the clothing style. The video concludes with a call to action for viewers to like, subscribe, and share the channel, and to reach out with any questions or comments. It also invites viewers to explore other programs on the channel.

Mindmap

Keywords

💡Midjourney

Midjourney is a text-to-image generator that uses diffusion models to create high-quality images from textual descriptions. It is central to the video's theme as it is the primary tool being discussed for generating continuous character images. For instance, the video mentions that 'Midjourney can generate high-quality images from text' and 'a prompt technique can be used to generate images of the same character from different angles and backgrounds using Midjourney.'

💡Diffusion Model

A diffusion model is a type of deep learning model used for generating images from textual descriptions. It is significant in the context of the video as it forms the basis of how Midjourney operates. The script refers to 'Midjourney and stable diffusion both using diffusion models to complete image generation.'

💡Prompt Techniques

Prompt techniques are methods used to guide the generation of images by providing specific textual instructions to the AI model. They are crucial for achieving the desired output in image generation. The video script discusses a new prompt technique discovered by a user named Julie, which allows for the generation of images of the same character from different angles and backgrounds.

💡Image Split

Image split refers to the division of a single image into multiple regions, each maintaining consistency in character features and background. It is a key concept in the video as it enables the creation of continuous character images. The script illustrates this with an example: 'In the prompt, adding the term image split into 2 can split a complete image into two areas, each with a consistent character face and background.'

💡Stable Diffusion

Stable Diffusion is another AI model used for generating images from text. It is mentioned in the video as an alternative to Midjourney, highlighting its capabilities and the availability of various models like SDXL 1.0. The script states, 'We can use the same prompt technique to generate continuous character images with stable diffusion models.'

💡Continuous Character Generation

Continuous character generation is the process of creating a series of images that depict the same character in different poses, angles, or backgrounds. It is the main focus of the video, demonstrating how to achieve this with Midjourney and stable diffusion. The script provides examples: 'We can design a set of corgi images with the scene divided into three parts, including different actions, or generate a set of character images that maintain a high degree of consistency.'

💡Segmind

Segmind is a new drawing platform recommended in the video for users without a Midjourney paid account. It allows users to log in with their Google accounts and provides a free trial with 100 points to use various stable diffusion models. The script mentions, 'I recommend a new drawing platform, Segmind, for those who do not have a Midjourney paid account.'

💡Tensor.art

Tensor.art is a platform mentioned in the video that offers free image generation services with a sufficient number of points and a rich selection of models. It is presented as a user-friendly alternative for generating images. The script states, 'You can also use the free Tensor.art platform to generate images, which provides enough points and a rich selection of models for more convenient use.'

💡Vector Illustrations

Vector illustrations refer to the creation of images using vector graphics, which are scalable and resolution-independent. They are used in various fields like publishing, media, and advertising. The video discusses using Midjourney to design professional vector illustrations, emphasizing its simplicity and effectiveness. The script provides a formula for creating vector-style icons: 'If you want to design vector-style icons, you can also refer to this prompt formula.'

💡Inpainting Tool

The inpainting tool is a feature in Midjourney that allows users to modify specific parts of an image, such as changing a character's clothing style. It is highlighted in the video as a way to enhance and customize the generated images. The script explains, 'We can also use Midjourney's newly added inpainting tool to replace the character's clothing, by selecting the character's body area and changing the original clothing style term.'

💡AnimateDiff Tool

AnimateDiff is a tool mentioned in the video that can combine images of different poses into an animation. It is presented as an additional feature for users to create dynamic content from static images. The script suggests, 'Using the AnimateDiff tool, you can merge images of different poses into an animation, which you can try on your own.'

Highlights

Midjourney can generate high-quality images through text prompts, but struggles with continuous character generation due to the lack of inpainting and controlnet plugins.

A prompt technique discovered by a user named Julie allows for the generation of images of the same character from different angles and backgrounds.

Adding 'image split into 2' in the prompt can divide a complete image into two consistent areas.

The split technique can be used to create a truly continuous image generation by dividing an image into multiple regions.

The method can be applied to create a series of images of corgis with different actions or human characters with high consistency.

Descriptive terms for character expressions, clothing, and background can be added to the prompt for enhanced image generation.

The technique can be used for anime character generation with the mode adjusted to 'niji5' and specific prompt terms.

Segmind is a new drawing platform offering free access with a Google account and a 100-point free allowance for new users.

SDXL 1.0 is a newly released model on Segmind that can generate a series of continuous character images from multiple angles.

Tensor.art is a free platform that provides ample points and a variety of models for image generation.

Stable diffusion's broader range of large models and LORA models allows for the generation of strongly continuous characters.

AnimateDiff is a tool that can merge images of different poses into an animation.

Midjourney can be used to design vector illustrations for various applications such as publishing, media, and advertising.

A simple formula is provided for designing vector illustrations with placeholders for the main subject and environment.

The formula allows for easy modification of the image's main content and background to generate corresponding images.

Changing prompt variables can add objects like dogs to the image, showcasing the simplicity of the method.

The image design style, color, and proportion can be altered to create diverse images based on the initial prompt.

Vector illustrations designed using this method can be sold on platforms like WireStock and AdobeStock, which support AI-generated images.

Midjourney's new inpainting tool can be used to replace the clothing style of a character in an image.

Casual Browsing

2024 midjourney设计人物角色详细教程，Midjourney 建立角色 + 分镜设计，全过程视频演示，Mj从零基础设计角色一致性，Mj连续人物和角色生成方法，如何为连续性人物添加不同的表情

2024-05-23 00:00:01

2024全新版本Midjourney绘图最详入门教程, 如何生成连续性人物，ChaoS参数详解 Seed参数 ,用Seed指令在Midjourney中进行连续创作，Mj以图生图，使用参数生成同一人物

2024-05-23 00:50:03

2024最新Midjourney以图生图零基础详细教程，生成人物统一连续/同一场景出现两个人物的方法，Mj如何垫图ai作图的方法,ai根据照片生成相似照片,用ChatGPT+AI绘画软件做故事视频赚钱

2024-05-19 02:00:02

midjourney v6 advanced tutorial ，MJ同一场景出现多个一致性角色的绘图方法，局部修改，改照片换脸，AI连续角色创作方法，AI绘图如何写提示语

2024-07-06 03:50:01

【新方法】免費生成AI網紅，超逼真零門檻｜比Midjourney更好用的AI生圖軟件｜Tensor Art｜角色一致｜AI influencer

2024-05-21 08:35:02