100%相似！Midjourney来制作属于自己的头像，通过InsigheFaceSwap来实现换脸，创作和垫图人物一模一样的头像图片，制作和垫图100%相似的头像

火箭學院 Rocket College
9 Jun 202308:03

TLDRIn this video, the creator introduces a process for generating personalized avatars using Midjourney and InsightFace. The tutorial begins with uploading a base image, then refining it with Midjourney's 'Imagine' feature to create a stylized portrait. Finally, InsightFace is used to swap faces, resulting in a highly detailed, custom avatar. The video showcases the ease and speed of this process, offering a creative way to produce unique digital portraits.

Takeaways

  • 🎨 The video introduces the process of creating personalized avatars using Midjourney and InsightFace.
  • 🌟 The presenter shares their experience with Midjourney's image generation capabilities.
  • 🔍 InsightFace is highlighted as a powerful, open-source face-swapping tool from MIT.
  • 📸 The process begins with uploading a base image to the server and ensuring it's correctly uploaded.
  • 🖌️ Midjourney is used to generate a new image based on a desired style and a set of keywords.
  • 💡 Keywords include elements like character, background, time, action, style, texture, and lighting.
  • 🛠️ The video provides a detailed guide on setting parameters for Midjourney, including model choice and style intensity.
  • 🔗 The base image's URL is used in conjunction with keywords to generate the new image.
  • 👤 InsightFace is added to Discord to perform the face swap, requiring a clear facial image of the base person.
  • 🔄 The face-swapping process involves saving the base facial features with InsightFace and then applying them to the generated image.
  • ⏱️ The entire process from image upload to final face swap can be completed in about 10 minutes.
  • 💌 The presenter encourages viewers to experiment with different styles and keywords for varied results.

Q & A

  • What is the main topic of the video?

    -The main topic of the video is about using Midjourney and InsightFace to create custom avatars by generating images based on a base image and then swapping the face to match a specific person's facial features.

  • Who is introduced as the model for the avatar creation in the video?

    -The model introduced for the avatar creation is Ayaka Miyoshi, a Japanese model and actress, recognized from Jay Chou's music video 'Won't Cry'.

  • How long does it take to create these avatar images?

    -It takes approximately 10 minutes to create these avatar images.

  • What is the first step in creating an avatar using Midjourney?

    -The first step is to upload the base image to the server and ensure it is successfully uploaded by hitting the enter key after adding the image to the chat box.

  • What is the theme chosen for the avatar creation in the video?

    -The theme chosen for the avatar creation in the video is 'Disney Pixar style'.

  • What are some of the keywords used for the Midjourney avatar creation?

    -Some of the keywords used include 'cool girl', 'golden hair', 'smiling and looking at you', 'Disney Pixar style', and 'super detail'.

  • What model and stylistic settings were chosen for the Midjourney creation?

    -The V5 model was chosen with a high stylization setting. The image ratio was set to 2:3, the base image weight was set to 1.5, and the image quality parameter was set to 2.

  • How does one use InsightFace to swap faces in images?

    -To use InsightFace for face swapping, one needs to join the InsightFace bot in Discord, save the base image person's facial features using the /saveid command, and then use the INSwapper app to swap the face in the previously created Midjourney image.

  • What is the source of the facial features used by InsightFace for the swap?

    -The facial features used by InsightFace for the swap come from a standard facial photo of the base image person, which should be uploaded using the /saveid command.

  • What is the final outcome of using both Midjourney and InsightFace?

    -The final outcome is an image where the generated avatar has the same facial features as the original base image person, in this case, Ayaka Miyoshi, but with the style and details generated by Midjourney.

  • What advice does the video creator give for those learning and practicing with Midjourney?

    -The video creator advises to try different styles of keywords for creation, as there can be millions of variations, and it may take multiple attempts to find a preferred result, similar to opening a blind box.

Outlines

00:00

🎨 Introducing Midjourney and InsightFace for Custom Avatars

This paragraph introduces the use of Midjourney for creating custom avatars through the process of 'image to image' generation. The creator shares their experience with using Midjourney and InsightFace to generate an avatar identical to the base image, using Japanese celebrity Miyoshi Ayaka as an example. The process involves uploading the base image, setting the desired avatar style (in this case, 'Disney Pixar style'), and using specific keywords to guide the Midjourney AI in creating the image. The creator emphasizes the ease and speed of the process, taking only 10 minutes to complete.

05:04

🤖 Detailed Steps for Creating and Swapping Faces with Midjourney and InsightFace

The second paragraph delves into the detailed steps of using Midjourney and InsightFace for face swapping in avatar creation. It begins with uploading the base image and using Midjourney to create a new avatar based on the selected style and keywords. The creator then explains how to use InsightFace, an advanced face recognition software, to replace the generated face with a real person's face, using Miyoshi Ayaka's face as an example. The paragraph includes instructions for saving the facial features with InsightFace and using it to swap faces in the generated avatar. The creator encourages viewers to experiment with different styles and keywords for a unique result.

Mindmap

Keywords

💡Midjourney

Midjourney is a term used in the context of an AI-based image generation tool that creates unique, stylistic portraits or images based on a user's input. In the video, it is the primary method for generating the base image or 'matte' from which further customization is done. The process involves uploading a base image and using keywords to guide the AI in creating a stylized version of the image, such as 'Disney Pixar style'.

💡InsightFace

InsightFace is an advanced, open-source face recognition software licensed by MIT. In the video, it is used to swap faces in images, taking the generated image from Midjourney and replacing the face with that of a specific person, such as the Japanese model and actress Ayaka Miyoshi. This process involves saving the facial features of the target individual and then applying them to the generated image.

💡Ayaka Miyoshi

Ayaka Miyoshi is a Japanese model and actress recognized in the video for her appearance in Jay Chou's music video 'Won't Cry'. She serves as the example of a celebrity whose face is used to replace the generated image's face using InsightFace, showcasing the software's capability to create realistic face swaps.

💡Face Swap

Face swap is the process of replacing the face in an image or video with another person's face. In the context of the video, it is achieved using the InsightFace software to create a realistic and detailed facial replacement, making the generated image appear as if it has the exact facial features of the original person in the base image.

💡Disney Pixar Style

Disney Pixar Style refers to the distinctive visual art style characteristic of Disney and Pixar animated films. In the video, it is used as a keyword to guide the Midjourney AI in creating a portrait with a specific aesthetic, suggesting elements like vibrant colors, expressive characters, and detailed backgrounds.

💡Keywords

In the context of the video, keywords are specific terms or phrases used to guide the AI in Midjourney to generate an image with desired characteristics. These can include descriptors of the subject, background, actions, style, materials, and lighting, which collectively shape the final image.

💡Discord

Discord is a communication platform used in the video to host and interact with the AI bots, such as Midjourney and InsightFace. It allows users to join servers and channels where they can upload images, exchange messages, and execute commands to generate content.

💡AI Bot

An AI bot, as discussed in the video, refers to an artificial intelligence application that performs specific tasks autonomously. In this case, the AI bots are Midjourney for image generation and InsightFace for face swapping. These bots are integrated into Discord servers and respond to user commands to create or modify images.

💡Image Generation

Image generation is the process of creating new images from existing ones or from scratch using AI algorithms. In the video, this is achieved through the Midjourney AI bot, which takes a base image and user-provided keywords to produce a stylized portrait.

💡Facial Features

Facial features refer to the specific details and characteristics of a person's face, including the eyes, nose, mouth, and overall shape. In the video, the term is used to describe the elements of Ayaka Miyoshi's face that are captured and applied to the generated image during the face swap process with InsightFace.

💡Stylized Portrait

A stylized portrait is an artistic representation of a person that deviates from a realistic depiction, often incorporating unique visual elements or techniques. In the video, the stylized portrait is created using Midjourney AI, with the aim of capturing the likeness of the base image while applying a specific artistic style, such as 'Disney Pixar'.

💡Image Quality Parameters

Image quality parameters are settings that influence the resolution, detail, and overall visual fidelity of an image. In the context of the video, these parameters are adjusted within the Midjourney AI to control the output quality of the generated image, with options like aspect ratio and image quality level.

Highlights

Introduction to Midjourney for creating unique avatars using AI.

Demonstration of creating an avatar identical to the base image using InsightFace AI.

Showcase of completed avatars featuring Japanese celebrity Ayaka Miyoshi.

Explanation of the 10-minute process to create these avatars.

Step-by-step guide on uploading the base image to the server.

Utilization of Midjourney AI to craft avatars in a 'Disney Pixar style'.

Selection of keywords such as 'cool girl', 'golden hair', and 'smiling'.

Translation of keywords into English and formatting for AI instructions.

Setting up AI parameters including model choice and style intensity.

Explanation of the /Imagine command and its usage with base image URL and keywords.

Adjustment of image ratio, base image weight, and quality parameters.

Process of selecting and refining the generated images.

Introduction to InsightFace, an advanced face-swapping AI software.

Instructions on adding the InsightFace bot to Discord and saving the base image's facial features.

Utilization of InsightFace for face-swapping to achieve a perfect match with the base image.

Discussion on the subtle yet significant changes after face-swapping.

Recommendation to experiment with different styles and keywords for diverse outcomes.

Encouragement for viewers to engage and ask questions for clarification.

Appreciation for viewership and a call to action for likes and subscriptions.

Create Stunning Music from Text with Brev.ai!

Turn your text into beautiful music in 30 seconds. Customize styles, instrumentals, and lyrics.

Try It Now

Casual Browsing

midjourney tutorial 2024最新midjourney头像制作教程，手把手叫你如何用Midjourney给头像换不同风格，如何垫图，以图生图，头像生成保持人物一致的方法，ai制作头像

2024-05-21 22:25:01

Midjourney以图生图的详细教程，零基础AI绘图，用上传的图片，让MJ生成相似的图片，如何用Midjourney生成相似图片以及参考图/垫图使用方法，好用的AI绘图工具推荐，如何一键生成连续人物

2024-05-21 22:20:01

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

2024-03-24 09:55:01

使用Ai绘图工具创作连续性人物和角色 - Leonardo, Midjourney, Stable Diffusion的操作技巧

2024-03-24 16:55:11

更逼真！Midjourney生成自己的专属头像，这次更像自己！绝密方法公开全流程演示教程

2024-03-26 02:20:02

最火的AI作图模型，这5款免费下载，含提示词，配合 Stable-diffusion 来制作高清大图吧！ | 零度解说

2024-04-14 22:10:00