【AIツール】Midjourney - ミッドジャーニーで写真を元に画像生成する方法。

HIROCODE.ヒロコード
7 Mar 202308:39

TLDRThe video script introduces an AI tool called Midjourney, which generates images based on specific keywords and reference photos. It explains the process of using Midjourney, including creating a Discord account, joining the Midjourney server, and executing commands to generate images. The script emphasizes the importance of using reference images and 'spells' (commands) to create images closer to one's vision. It also discusses the free and paid plans, commercial use of generated images, and various parameters that can be used to improve the quality of the generated images. The video aims to showcase how easily high-quality images can be created with AI, hinting at the future where using AI for work becomes commonplace.

Takeaways

  • 🌟 Introduction to AI tool, Midjourney, which generates images based on specific keywords and can also incorporate reference photos for more accurate results.
  • 📸 Midjourney allows users to upload images to create images closer to their envisioned concepts, overcoming the limitations of text-only descriptions.
  • 💡 The process of using Midjourney involves creating a Discord account, joining the Midjourney server, entering the command, and generating images.
  • 💰 Midjourney offers a free plan for generating up to 25 images, with paid plans available for more extensive use, including commercial use of the generated images.
  • 🎨 The quality of generated images can be influenced by the background color of the reference images, so it's important to choose appropriate reference images.
  • 🔗 Users can upload images directly to Discord and use the image URL in combination with 'spells' (commands) to generate new images.
  • 📝 The 'spells' or commands used in Midjourney can significantly alter the output, so experimenting with different combinations is recommended.
  • 🔄 Midjourney provides options to refine the generated images, such as high-quality enhancement (U1-U10) and regenerating images with the same 'spell' (recycling mark).
  • 🔍 Users can reference other people's posts on Discord to find successful 'spells' and generate images closer to their desired outcome.
  • 🔧 Parameters like -AR for aspect ratio and - for excluding specific keywords can be used to fine-tune the image generation process.
  • 🌿 The presenter experimented with combining multiple reference images and keywords, such as a portrait and a plant photo, to generate a composite image.

Q & A

  • What is the AI tool introduced in the script?

    -The AI tool introduced in the script is called Midjourney.

  • How does Midjourney generate images?

    -Midjourney generates images based on specific keywords provided by the user. It can also use reference images to create images closer to the user's imagination.

  • What is the significance of the '咒文' (incantation or command) in Midjourney?

    -The '咒文' or command is a set of instructions or keywords that the user provides to Midjourney to generate a specific type of image. The choice of words can greatly affect the resulting image.

  • What is the basic process of using Midjourney?

    -The basic process involves creating a Discord account, joining Midjourney through an invitation, entering the appropriate room, executing commands, and generating images.

  • What are the limitations of the free plan on Midjourney?

    -The free plan on Midjourney allows for the generation of up to 25 images. To generate more, a paid plan is required.

  • What is the Basic Plan on Midjourney and how much does it cost?

    -The Basic Plan on Midjourney costs $10 per month and allows for the generation of up to 200 images.

  • What are the commercial usage rights for images generated on Midjourney?

    -Images generated on Midjourney are not allowed for commercial use by default. However, once a user subscribes to a paid plan, commercial usage rights are granted.

  • How can the background color of a reference image affect the generated image?

    -The background color of a reference image can influence the final result, as seen in the script where a transparent background (PING) and a white background (JPEG) produced different results.

  • What is the purpose of the parameters like -AR, -NOT, and others during image generation?

    -Parameters like -AR are used to modify the aspect ratio of the generated image, while -NOT allows for the exclusion of specific keywords. These parameters help in fine-tuning the image generation to better match the user's requirements.

  • How can looking at other people's posts on Discord improve the image generation process?

    -By examining other people's posts and the keywords they used, users can gain insights and use similar commands to generate images closer to their desired outcome.

  • What was the result of using multiple reference images in the script?

    -Using multiple reference images resulted in a generated image that reflected elements from both photos, showing a combination of the input images and producing a more complex result.

Outlines

00:00

🖼️ Introduction to AI Image Generation with Midjourney

This paragraph introduces the use of AI tool Midjourney for image generation based on specific photographs. It explains that while Midjourney typically generates images from text keywords, there are limitations to expressing detailed personal visions with text alone. The video aims to demonstrate how to generate images closer to one's own imagination by combining text with reference photo data. It also mentions the 'incantations' or commands used in image generation and their impact on the resulting images.

05:01

📸 Preparing and Uploading Reference Photos

The speaker discusses the process of preparing and uploading reference images for Midjourney to generate images that closely match one's vision. It highlights the importance of selecting appropriate images and the impact of background color on the generation results. The speaker shares their trial and error experience and emphasizes the need to use images that are publicly acceptable. The paragraph outlines the steps for uploading images to Discord and the initial interaction with the Midjourney bot.

🔍 Adjusting and Refining Generated Images

This section delves into the adjustments and refinements that can be made to the generated images. It explains the use of various buttons available after image generation, such as high-quality enhancement (U1 to UFO) and regenerating images based on the same incantation (recycling mark). The paragraph cautions about the consumption of generation attempts when using these buttons and encourages viewers to experiment with different keywords and incantations to achieve desired results.

🌟 Advanced Techniques and Multiple Photo Usage

The speaker introduces advanced techniques for image generation, including the use of specific parameters like -AR for aspect ratio adjustment and - to exclude certain keywords. It also suggests referencing other people's posts for inspiration and improving the quality of images by incorporating high-quality and beautiful keywords. The paragraph concludes with a demonstration of generating images using multiple reference photos, showcasing the versatility and potential of Midjourney in creating complex and detailed images.

🚀 Reflecting on AI's Role in Image Creation

In the final paragraph, the speaker reflects on the ease and quality of image generation using AI and Midjourney, emphasizing the transformative impact of technology. It suggests that the use of AI in various tasks will become commonplace in the near future. The speaker encourages viewers to explore AI, especially those who have never used it before, and invites feedback and engagement from the audience.

Mindmap

Keywords

💡AI工具ミッドジャーニー (AI tool Midjourney)

AI tool Midjourney is a service that generates images based on specific keywords. It is an AI platform that can create images not only from text descriptions but also by incorporating reference images, allowing users to produce visuals closer to their imagined concepts. In the video, the speaker discusses how to use Midjourney to generate images that align more closely with their personal visions by combining text and reference images.

💡画像生成 (Image generation)

Image generation refers to the process of creating visual content using AI algorithms. In the context of the video, it involves using Midjourney to produce images based on textual descriptions and reference photos. The goal is to generate images that closely match the user's intended vision, overcoming the limitations of expressing complex ideas through text alone.

💡コマンド (Command)

In the context of the video, a command refers to the specific instructions or 'incantations' used in Midjourney to generate images. These commands can include text descriptions, reference image URLs, and various parameters that guide the AI in creating the desired output. The choice of commands can significantly impact the resulting image, making them a crucial aspect of using the tool effectively.

💡Discord

Discord is a communication platform where users can create and join various 'rooms' or servers to interact with others. In the video, the speaker describes using Discord to access Midjourney, where they can upload reference images and execute commands to generate images. Discord serves as the interface for interacting with the AI tool and sharing results.

💡無料プラン (Free plan)

The free plan is an option offered by Midjourney that allows users to generate a limited number of images without any cost. It serves as an entry point for users to experience the AI tool and its capabilities before deciding to upgrade to a paid plan for more extensive usage.

💡有料プラン (Paid plan)

A paid plan is a subscription model offered by Midjourney that provides users with the ability to generate a larger number of images compared to the free plan. These plans typically come with additional features and capabilities, such as higher image quality and the option to use generated images for commercial purposes.

💡商用利用 (Commercial use)

Commercial use refers to the application of a product, service, or content in a business context, typically for profit-making purposes. In the context of the video, it discusses the limitations around the commercial use of images generated by Midjourney, noting that such usage is not allowed under the free plan but becomes permissible with a paid plan subscription.

💡参考画像 (Reference image)

A reference image is a visual example that users provide to the AI tool to guide the generation process. By including a reference image along with textual descriptions, users can direct the AI to create images that more closely align with their intended concepts or styles.

💡呪文 (Incantations)

In the context of the video, 'incantations' is a playful term used to describe the specific commands or text inputs that users provide to Midjourney to generate images. These incantations combine keywords, reference image URLs, and other parameters to guide the AI in producing the desired visual outcomes.

💡パラメーター (Parameters)

Parameters are specific settings or options that users can adjust within the AI tool to influence the image generation process. These can include aspects like aspect ratio, exclusion of certain keywords, or the inclusion of specific styles or qualities that can alter the final output.

💡複数の写真 (Multiple photos)

Refers to the use of more than one reference photo in the image generation process. By incorporating multiple photos, users can provide the AI with a broader range of visual cues and styles, potentially leading to more complex and detailed images that reflect a combination of the provided references.

Highlights

Introduction to AI tool Midjourney for generating images based on specific keywords and photographs.

Midjourney generates images based on text keywords, but sometimes there are limitations in expressing detailed ideas with text alone.

Combining text with reference photos can lead to generating images closer to one's imagination.

Explanation of the command '咒文' (incantation) used in image generation and how it affects the outcome.

Step-by-step guide on using Midjourney, including creating a Discord account, joining the Midjourney server, and executing commands.

Information on the free and paid plans available for Midjourney, including the number of images that can be generated and the cost.

Instructions on how to upload reference images and use them in the image generation process.

The impact of background color in reference images on the generation results and the possibility of changing it with '咒文'.

Demonstration of the actual image generation process, including selecting the 'スラッシュイマジン' prompt and inputting the '咒文'.

Explanation of the waiting time for image generation and the factors that affect it.

Showcase of the final generated image and comparison with the reference photo to demonstrate the effectiveness of the process.

Discussion on how to refine the '咒文' to achieve better results if the initial image does not match the desired outcome.

Introduction to additional parameters that can be used during image generation, such as aspect ratio and exclusion of specific keywords.

Advice on using other people's successful '咒文' as a reference to improve one's own image generation.

Example of using multiple reference images to generate a more complex and detailed image.

Reflection on the ease of generating high-quality images with AI and the potential future of AI in various fields of work.

Encouragement for those who have never used AI before to try Midjourney and experience the capabilities of AI technology.

Conclusion and appreciation for watching the video, with a call to action for likes and channel subscriptions.

Casual Browsing

【誰でもできる！】ChatGPTとMidjourneyで超リアルな画像生成をする方法。

2024-04-16 02:00:01

画像を元に新たな画像を生成するAI「Stable Diffusion Reimagine」が便利！【Clipdrop】

2024-04-21 15:45:01

画像生成AIにモデルを追加する方法【AUTOMATIC1111 Stable Diffusion web UI】

2024-03-30 15:00:01

画像生成AIのプロンプトをChatGPTで簡単に作る方法【Midjourney、Stable Diffusion対応】

2024-04-15 04:55:01

【stable diffusion】元画像のままアップスケールする方法

2024-04-21 15:30:04

WebUI ForgeでPhotomakerを使う方法を紹介します【Stable Diffusionで画像と同じ人物を生成する！】

2024-04-16 14:20:02