動漫真人化! 葬送的芙莉蓮 Midjourney教學 Prompt影片化 AI sora

蘋果妹
16 Mar 202414:17

TLDRIn this tutorial, the presenter guides viewers on how to use Midjourney to transform their favorite anime characters into realistic images. The process involves three main steps: using the /describe tool to generate initial prompts, refining the prompts to create a realistic version of the character, and making final adjustments to details. The video uses the character Faerun from the anime 'Fulian' as an example, discussing character traits and how to avoid common pitfalls like generating overly thin characters. The presenter also covers the use of specific settings like Niji mode for anime-style images and the --s parameter to adjust creativity levels. The tutorial is rich with practical tips, such as using 'No anime' in prompts to avoid unrealistic results, and the --Repeat function for sending the same prompt multiple times. The presenter also touches on the challenges of rendering complex costumes and suggests using Stable Diffusion for highly detailed images. The video concludes with an invitation for viewers to share their experiences or request anime characters they'd like to see.

Takeaways

  • 📚 **Introduction to Midjourney**: The video teaches how to use Midjourney to create real-life versions of anime characters.
  • 🎨 **Character Selection**: Faerun from the anime 'Fulian' is chosen for demonstration due to his simple design, making him an ideal candidate for beginners.
  • 🧙 **Character Analysis**: Faerun is described as a non-stereotypical girl with a round face, emotional personality, and a rare magic wizard.
  • 📈 **Three-Step Process**: The process involves using cartoon pictures to reverse the Prompt, refining the real-life Prompt, and local detail refinement.
  • 🔍 **/describe Tool**: This tool analyzes images to generate prompts, which are then tested in Niji mode for anime-style pictures.
  • 🔧 **Prompt Refinement**: The video emphasizes the need to refine prompts to avoid generating unrealistic or unwanted character traits.
  • 🎭 **Emphasizing Realism**: The use of descriptors like 'Realistic', 'Life Action', and 'Real Human' is crucial to achieve a non-anime, life-like result.
  • 👚 **Clothing Description**: Describing clothing accurately is challenging; using simple terms or consulting external sources like ChatGPT can be helpful.
  • 🧞 **Handling Details**: For complex costumes or hairstyles, emphasizing key features and using tools like --Repeat can improve results.
  • 🖌️ **Photoshop Touch-ups**: Final adjustments using Photoshop can correct minor issues and enhance the final image.
  • ⚙️ **Technical Adjustments**: Using the --s parameter to adjust the Stylize threshold can help in fine-tuning the creativity level of the generated images.

Q & A

  • What is the main topic of the video?

    -The main topic of the video is teaching viewers how to use Midjourney to create real-life versions of anime characters, specifically focusing on characters from the anime 'Fulian'.

  • What is the first step in using Midjourney for character creation?

    -The first step is to choose a character and use the /describe tool to find inspiration for the Prompt by uploading a reference picture.

  • Why is Niji mode used in the process?

    -Niji mode is a model specially used to generate anime pictures, which is suitable for the task of bringing anime characters to life.

  • What is the significance of switching back to V6 setting before generating a real-life version of the character?

    -Switching back to V6 setting ensures that the generated image is closer to a real-life representation rather than an anime style, which is the default in Niji mode.

  • How does the video script address the challenge of Midjourney generating thin characters?

    -The script suggests using specific words like 'A Chubby Girl' and 'Slightly Round Face' to counteract Midjourney's default tendency to generate thin characters.

  • What is the purpose of adjusting the --s threshold in Midjourney?

    -Adjusting the --s threshold, or Stylize threshold, controls the level of creativity and adherence to the Prompt. A higher value increases creativity but may result in more confusion, while a lower value makes the output more conservative and closer to the given instructions.

  • How does the video script suggest refining the character's appearance in the Prompt?

    -The script suggests removing irrelevant words from the Describe output and focusing on key characteristics of the character, such as 'White Conservative Dress' and 'High Necked' to suppress unwanted features like large cleavage.

  • What is the role of the 'No anime' command in the Prompt?

    -The 'No anime' command is used to avoid generating anime-style images, ensuring that the output is more aligned with real-life or live-action representations.

  • How does the video script handle complex costumes or fantasy elements in character design?

    -For complex costumes or fantasy elements, the script suggests using the Describe tool to generate Prompt words and, if necessary, consulting ChatGPT for single-word descriptions to simplify the process.

  • What tool is mentioned for making minor adjustments to specific parts of the generated image?

    -The 'Very Region' tool is mentioned for making minor adjustments to specific parts of the generated image, such as modifying the hair or the edge of clothes.

  • What is the final step the presenter takes to improve the generated image?

    -The final step the presenter takes is to use Photoshop for further refinement, adjusting details like hair volume, bangs, and facial features to achieve a more satisfactory result.

  • Why is Stable Diffusion mentioned as a more suitable tool for high-detail character costumes?

    -Stable Diffusion is mentioned as more suitable for high-detail character costumes because it can handle complex textures and patterns better than Midjourney, which might struggle with intricate details.

Outlines

00:00

🎨 Introduction to Midjourney: Bringing Anime Characters to Life

The video begins with an introduction to using Midjourney, a tool for creating art, to bring anime characters to life. The presenter chooses Faerun from the anime Fulian as an example and discusses the importance of selecting the right character and understanding their background for accurate representation. The process involves three steps: using cartoon pictures to reverse-engineer prompts, refining the real-life prompt, and local detail refinement. The video also provides guidance for beginners, including how to use the /describe tool for inspiration and switching to Niji mode for anime-style images. The presenter emphasizes the need to avoid certain character traits and to focus on the character's personality and abilities when crafting prompts.

05:02

🧙‍♂️ Crafting a Realistic Faerun: From Anime to Real Life

The second paragraph delves into the process of transforming the anime character Faerun into a realistic, live-action version. The presenter discusses the challenges of generating the correct body type and clothing, and how to refine the prompt to achieve the desired outcome. The video covers the use of specific adjectives to convey emotions, personality, and physical attributes. It also explains the use of the --s parameter to adjust the level of creativity in the generated images and the importance of avoiding unwanted elements with the 'No' command. The presenter shares their satisfaction with the final image and discusses the process of fine-tuning it using Photoshop.

10:03

🧝‍♀️ Portraying Fulian: Simplifying Complexity for Realism

The third paragraph focuses on the character Fulian, an elf from the same anime. The presenter describes the initial challenges in portraying Fulian's complex costume and suggests using AI tools like ChatGPT to describe the outfit. The video outlines the steps for refining the prompt, emphasizing the importance of specifying details like the 'Japanese Twintails Hairstyle' and using the 'No anime' command to avoid unrealistic portrayals. The presenter also introduces the --Repeat tool for sending the same prompt multiple times and the Very Region tool for modifying specific parts of the generated image. The video concludes with a discussion on the limitations of Midjourney for highly detailed costumes and the presenter's personal adjustments to the final image using Photoshop.

Mindmap

Keywords

💡Midjourney

Midjourney is a term used in the context of the video to refer to a specific AI drawing tool or platform that allows users to generate images based on textual prompts. It is central to the video's theme as the entire tutorial is about how to use this tool to create real-life representations of anime characters.

💡Anime Character

An anime character refers to a character from Japanese animation, or anime. In the video, the focus is on bringing such characters to life through the Midjourney tool, specifically using characters from the anime 'Fulian' as examples.

💡Prompt

A prompt in the context of AI drawing is a set of instructions or a description that guides the AI in generating an image. The video discusses how to craft effective prompts to achieve desired results with the Midjourney tool.

💡Niji mode

Niji mode is a specific setting within the Midjourney tool designed to generate anime-style images. It is mentioned in the video as a necessary step when attempting to create anime character representations.

💡Real-life Version

Refers to the process of transforming a fictional or animated character into a more lifelike or realistic representation. The video provides a step-by-step guide on achieving this with the Midjourney tool.

💡/describe tool

The /describe tool is a feature within Midjourney that analyzes an uploaded image and generates a set of prompts based on the image's content. It is used in the video to find inspiration for creating prompts.

💡Stylize Threshold (--s)

The Stylize Threshold, denoted by --s followed by a number, is a parameter in Midjourney that adjusts the level of creativity and adherence to the prompt. A higher value increases creativity but may result in less accurate representations, while a lower value makes the output more conservative and closer to the prompt.

💡Harry Potter

Harry Potter is used as a thematic reference in the video to help generate images that resemble scenes from the Harry Potter movies. It serves as an example of a real-life, magical setting that can be used to inform the style of the generated images.

💡Faerun

Faerun is a specific character from the anime 'Fulian' that is used as a primary example throughout the video. The character's traits, such as being a magic wizard and having a round face, are detailed to demonstrate how to create a prompt that accurately represents the character.

💡Fulilian

Fulilian is another character from 'Fulian' that is discussed in the video. The tutorial covers how to handle the complexity of her costume and posture using the Midjourney tool, emphasizing the challenges and solutions in creating a real-life version of her.

💡Photoshop

Photoshop is a popular image editing software mentioned in the video for refining the final generated images. It is used to make adjustments such as trimming details or changing hair texture to better match the desired character representation.

Highlights

The video teaches how to use Midjourney to bring anime characters to life.

Demonstrates using the anime character Fulian from the cover.

Three pictures of the character are used to increase the probability of success.

The process is divided into three steps: reverse Prompt using cartoon pictures, refining the real-life Prompt, and local detail refinement.

Introduction to the character Faerun, a 16 or 17-year-old girl with a round appearance and a theatrical personality.

Faerun is a rare magic wizard, described as a super genius.

Avoid making Faerun look thin or yandere in the Prompt.

Using the /describe tool to find inspiration for the Prompt.

Switching to Niji mode, a model for generating anime pictures.

Refining the Prompt by removing irrelevant words and focusing on character traits.

Describing Faerun's appearance with specific adjectives to balance the image generation.

Adding action, emotion, and environment descriptions to the Prompt for a more dynamic image.

Using the --s parameter to adjust the Stylize threshold for creativity.

Choosing a simpler image of Fulilian in pajamas for easier Prompt generation.

Fulilian is described as a thousand-year-old elf with a misanthropic personality.

Using the No command in the Prompt to avoid unwanted elements in the generated image.

Challenge of describing complex costumes, suggesting the use of ChatGPT for assistance.

Emphasizing the realism in the Prompt with terms like Realistic, Life Action, and Real Human.

Using the --Repeat tool for sending the same prompt multiple times to find the best result.

Utilizing the Very Region tool for modifying specific parts of the generated image.

Final adjustments made using Photoshop to achieve the desired character appearance.

Suggestion to use Stable Diffusion for high-detail characters.

Casual Browsing

再見了Midjourney｜Canva AI完整攻略教學，比Midjourney/Stable Diffusion更加實用的人工智能 #midjourney #canva #chatgpt

2024-04-29 22:25:02

【AI動画】進化したRunwayでAI映画＆アニメを作る方法！

2024-05-17 23:25:03

出神入化 AI 繪圖 Midjourney 參數解析！

2024-04-30 03:30:00

【保姆級實操】AI自動化做漫畫小說頻道，AI賺錢最新玩法 ，1萬美金/月 ｜Earn Money With AI By Creating Comic Channel

2024-05-16 22:10:03

【手把手教学】我用ChatGPT+Midjourney零成本打造出了一个线上服装生意｜全自动化赚钱流程公开，人人都可以做 #ai #chatgpt #midjourney

2024-05-03 08:40:01