【Stable Diffusion】ポーズの呪文集プロンプト一覧を紹介！

AIジェネ【AIイラスト生成の情報発信】
5 Aug 202311:12

TLDRThe video provides an in-depth guide on using 'pose spells' to generate a variety of poses in Stable Diffusion, a popular AI image generation tool. It explains the importance of entering the correct pose spell for desired outcomes, such as standing, sitting, walking, or lying down, and how to enhance the probability of generating specific poses. The transcript covers tips like hiding hands to avoid AI's difficulty with fingers, adjusting aspect ratios for better image balance, and prioritizing certain spells at the beginning of prompts for higher success rates. It also suggests using numerical values to adjust the spell's influence and recommends not exceeding 1.3 to maintain image quality. The video concludes with advice on using 'open pose editor' for more complex or unlisted poses, encouraging viewers to experiment with the spells introduced to create their ideal images.

Takeaways

  • 🔍 Use specific 'pose spells' in Stable Diffusion to increase the variation and accuracy of generated poses.
  • 🧍 If you want a standing pose, use the 'standing' spell for better results.
  • 🦵 To generate a pose with one leg up, add 'standing, raise one leg' to the prompt.
  • 🤗 For folded or crossed arms, which can help hide hands, use 'fold-arms' or 'crossed arms'.
  • 🤲 Adding 'arms behind back' or 'hand between legs' can be a good way to pose the character's hands.
  • 🚗 For driving poses, include 'hold the steering wheel of a car' in the prompt.
  • 👆 To indicate reaching out or a specific hand gesture, use 'reaching out' or 'put up index finger'.
  • 🧘‍♀️ For sitting poses, use 'sit' or 'sitting', and consider vertical image sizes for balance.
  • 🏃‍♂️ To generate walking or running poses, add 'walking' or 'running' to the prompt.
  • 🛌 For lying down or sleeping poses, adjust the aspect ratio to vertical or horizontal accordingly.
  • 📐 Changing the aspect ratio can significantly affect the success of generating the desired pose.

Q & A

  • What is the importance of using pose spells in Stable Diffusion?

    -Pose spells in Stable Diffusion are crucial for generating specific poses. Without the correct pose spell, the desired pose may not appear, or it may take a long time to find the right spell. Using pose spells increases the probability of generating the intended pose and helps in creating a more varied range of poses.

  • How can I increase the chances of generating a standing pose in Stable Diffusion?

    -To increase the chances of generating a standing pose, use the standing spell. Additionally, if you want the character to stand with one leg up, add 'standing, raise one leg' to the prompt.

  • What is a recommended method to hide hands in AI illustrations?

    -A recommended method to hide hands is to use the 'fold-arms' or 'crossed arms' spell. This is particularly useful because AI illustrations often struggle to create detailed and beautiful fingers.

  • How can I generate an image of a person with their hands behind their back?

    -To generate an image with hands behind the back, add 'arms behind back' to the prompt. This is another method to hide the hands, which can lead to better image generation.

  • What should I add to the prompt if I want the character to put their hands on their hips?

    -To generate a pose with hands on hips, add 'put hands hip'. However, it's more common for both hands to be placed on the waist, so if you want one hand on the waist, you might need to specify 'one hand on hip'.

  • How can I generate a pose where the character is holding the steering wheel of a car?

    -To generate a pose with the character holding the steering wheel, enter 'hold the steering wheel of a car'. It's noted that this pose is less likely to be generated in a portrait orientation and is better suited for square or horizontally long images.

  • What aspect ratio is recommended for generating a standing or sitting pose?

    -A vertical aspect ratio is recommended for generating standing or sitting poses as it tends to be easier to generate in this format.

  • How can I adjust the influence of a pose spell in Stable Diffusion?

    -You can adjust the influence of a pose spell by adding a numerical value after the spell, with a maximum recommended value of 1.3. This helps to strengthen the spell's influence without deteriorating the overall image quality.

  • Why is it suggested to place the pose spell at the beginning of the prompt in Stable Diffusion?

    -Placing the pose spell at the beginning of the prompt increases the likelihood of generating the desired pose because Stable Diffusion tends to prioritize spells that are listed at the top of the prompt.

  • What is the suggested method to generate an image with an ideal pose if it's not covered in the provided spells?

    -If the desired pose is not covered by the provided spells or cannot be generated, using 'open pose' or 'open pose editor' can be a faster method to specify and generate the ideal pose.

  • How can changing the aspect ratio help in generating the desired pose?

    -Changing the aspect ratio to vertical, square, or landscape can help in generating the desired pose. Different poses may be easier to generate in different aspect ratios, so trial and error with various ratios is recommended.

  • What is the recommended approach if the generated pose does not meet the expectation?

    -If the generated pose does not meet expectations, it is recommended to change the aspect ratio, adjust the numerical value to influence the spell's strength, and prioritize the pose spell by placing it at the beginning of the prompt.

Outlines

00:00

🎭 Introduction to Pose Spells for Image Generation

This paragraph introduces the concept of 'pose spells' used in stable diffusion for generating images with specific poses. It emphasizes the importance of using the correct spell to increase the likelihood of achieving the desired pose. The video offers a collection of pose spells for various stances, including standing, sitting, and lying down, and provides tips for hiding hands or creating more natural-looking images. It also discusses the influence of image aspect ratio on pose generation and suggests entering a numerical value to adjust the spell's influence for better results.

05:03

🖼️ Generating Ideal Poses with Aspect Ratio and Spell Adjustments

The second paragraph delves into the practical application of pose spells and offers troubleshooting tips for when the desired pose isn't generated. It suggests changing the aspect ratio to a vertical, square, or landscape orientation depending on the pose. The paragraph also recommends adjusting the numerical value after the spell to enhance its effect, with a caution not to exceed 1.3 to avoid quality deterioration. Furthermore, it advises prioritizing the pose spell at the beginning of the prompt for better results and using open pose or open pose editor for more control over the generated pose.

10:04

📈 Tips and Tricks for Successful Image Pose Generation

The final paragraph summarizes the key points discussed in the video. It reiterates the effectiveness of changing the aspect ratio and adjusting the spell's influence to generate the desired pose. The paragraph also encourages viewers to bring the pose spell to the front if they want it to be prioritized in the image generation process. It concludes by inviting viewers to subscribe to the channel for more content on generated AI and thanking them for watching.

Mindmap

Keywords

💡Stable Diffusion

Stable Diffusion is a term referring to a type of generative model used in AI, specifically for creating images from textual descriptions. In the context of the video, it is the technology that the user interacts with to generate various poses for images. It is central to the video's theme as it is the tool that enables the creation of the desired poses.

💡Pose Spells

Pose Spells are specific textual prompts or commands used to guide the Stable Diffusion model to generate images with particular poses. They are a key concept in the video as they are the method by which users can influence the outcome of the image generation process, aiming to achieve the desired pose in the final image.

💡Image Generation

Image Generation is the process of creating visual content from textual descriptions using AI models like Stable Diffusion. It is the main theme of the video, as the entire script revolves around how to effectively use Stable Diffusion to generate images with specific poses.

💡Aspect Ratio

Aspect Ratio refers to the proportional relationship between the width and the height of an image. In the video, it is discussed as a factor that can influence the success of generating a desired pose, with different poses being easier to achieve with certain aspect ratios.

💡Numerical Value

Numerical Value, in the context of the video, is a weight assigned to a pose spell to increase its influence on the image generation process. It is used to fine-tune the strength of a particular pose spell, with a recommended maximum value of 1.3 to maintain image quality.

💡Open Pose Editor

Open Pose Editor is a tool or feature that allows users to specify or edit poses more directly when using Stable Diffusion. It is mentioned as a method to potentially speed up the process of generating an ideal pose, especially when the desired pose is not among the standard options.

💡Portrait

Portrait in the video refers to a specific orientation of an image where the height is greater than the width, typically used for vertical compositions. It is discussed in relation to the aspect ratio and its effect on the generation of certain poses.

💡Landscape

Landscape is an image orientation where the width is greater than the height, often used for horizontal compositions. It is mentioned in the context of aspect ratios that are more conducive to generating poses where the character is lying down or sitting.

💡Square

Square, in the context of the video, refers to an image with equal width and height dimensions. It is one of the aspect ratios discussed as potentially beneficial for generating certain types of poses, such as those where the character is sitting or lying face down.

💡AI Gene

AI Gene is likely a reference to the channel or the entity responsible for creating and disseminating information about AI-generated content. It is mentioned at the end of the video as a source for more information and content related to AI and image generation.

💡Trial and Error

Trial and Error is a methodological approach where one experiments with different variables to find a successful outcome. In the video, it is suggested as a technique for users to employ when trying to generate the ideal pose, by adjusting aspect ratios and other parameters.

Highlights

Stable diffusion often requires the correct pose spell to generate desired poses.

Adding a standing spell increases the probability of generating a standing pose.

Combining spells like 'standing, raise one leg' can generate more complex poses.

Using 'fold-arms' or 'crossed arms' can help hide hands, which are difficult for AI to create.

Adding 'arms behind back' is a recommended method for hiding hands in image generation.

Including 'hand between legs' along with 'sit' or 'sitting' can improve pose generation.

Adding 'put hands hip' or 'one hand on hip' can help generate poses with hands on hips.

Using 'own hands together' can be challenging due to finger rendering complexities.

Entering 'hold the steering wheel of a car' can generate poses with the character driving.

Aspect ratios can influence pose generation, with vertical being more suitable for standing or sitting poses.

Adjusting the spell's influence with a numerical value (e.g., run1.3) can help achieve the desired pose.

The recommended value for adjusting the spell's influence is up to 1.3 to maintain image quality.

Placing the pose spell at the beginning of the prompt can increase the likelihood of generation.

Specifying poses with 'open pose' or 'open pose editor' can shorten the time to generate an ideal pose.

Changing the aspect ratio to vertical, square, or landscape can help generate poses that are difficult to achieve.

Using 'lie down', 'sleep', or 'lie face down' can generate poses for lying or sleeping characters.

Inserting 'walking', 'running', or 'jumping' can generate dynamic poses for movement.

Using 'kneel', 'wariza', 'seiza', or 'lotus position' can generate poses for sitting or kneeling.

The video provides a collection of pose spells and tips for successfully generating poses in stable diffusion.