How to use Midjourney Image to Image (Best Image Prompting Methods!)

Wade McMaster - Creator Impact
10 Aug 202307:28

TLDRThe video provides a comprehensive guide on how to effectively use image prompts in Midjourney, an AI art generation tool. It starts by explaining the basic method of using an image URL in the prompt. The video then offers alternative solutions for obtaining image URLs, such as using Discord or postimage.cc, and demonstrates how to adjust the 'Image Weight' to control the influence of the image on the generated art. It also introduces advanced techniques like 'chain prompting,' where an AI-generated image is used in a prompt to create another image, and 'blend,' which combines multiple images into one. The video concludes with the 'describe' function, which generates descriptive prompts based on an uploaded image. The host encourages viewers to share their own tips in the comments and to like the video if they found it helpful.

Takeaways

  • 🌐 Use a URL from an image to prompt Midjourney by adding it to your prompt.
  • 📤 Copy an image URL from Discord by uploading an image and copying the link address.
  • 📱 Similar process on a phone involves finding and copying the image URL.
  • 🖇️ Postimage.cc is a tool to upload an image and get a direct link for use in prompts.
  • 🔄 Image Weight (IW) can control how strongly Midjourney references the image in its prompt.
  • 🎭 Remix mode allows re-rolling and adjusting the image weight for different references.
  • 🖼️ Upscaled images can be used as prompts to create detailed outputs.
  • 🔗 Chain prompting involves using an AI-generated image in a prompt to generate another image, and so on.
  • 📈 Blend function combines multiple images into one prompt without additional text.
  • 🎨 Variations can be made on a single image to explore different styles.
  • 📝 Describe function generates descriptive prompts based on an uploaded image.
  • 🔄 Combining 'describe' with an image prompt can yield unique and creative results.

Q & A

  • What is the most basic way to use an image in Midjourney?

    -The most basic way to use an image is to insert the image URL into your prompt and add a bit of information to the end, then hit enter to get a generation inspired by your image.

  • How can you get an image URL to use in Midjourney if you don't have access to a website?

    -You can upload an image to a platform like Discord, then right-click the image and copy the link address. Alternatively, you can use a service like postimage.cc to upload the image and get a direct link.

  • What does the Image Weight parameter do in Midjourney?

    -Image Weight (IW) allows you to control how strongly Midjourney references the image in its prompt. The default is one, but you can increase it to a maximum of 2 for a closer reference.

  • How can you reference an image less in your prompt?

    -You can reduce the Image Weight to a lower value, such as 0.5, to make the image reference less prominent in the generated output.

  • What is chain prompting and how is it done in Midjourney?

    -Chain prompting is the process of using an AI-generated image in your prompt to generate another image, then using that new image in another prompt, and so on, to evolve your art from image to image.

  • Can you use multiple images in a single prompt in Midjourney?

    -Yes, you can include multiple image URLs in a single prompt to create a blend of styles and elements from different images.

  • What is the 'blend' function in Midjourney and how does it work?

    -The 'blend' function allows you to upload and combine multiple images into one prompt. You can upload up to five images and blend them, but you cannot add a prompt directly with blend; you need to use remix mode and re-roll to add a prompt.

  • How can you use the 'describe' function in Midjourney to enhance your prompts?

    -The 'describe' function generates four descriptive prompts based on a single uploaded image. You can use these prompts as they are, edit them, or imagine them one by one to see what Midjourney produces.

  • What is remix mode in Midjourney and how does it affect the image prompting?

    -Remix mode allows you to make additional changes to your prompt after the initial generation. With remix mode on, you can re-roll the image weight, add more descriptive elements, or combine describe functions with image prompts for more unique results.

  • How can you use variations on an image to create different outcomes in Midjourney?

    -With remix mode on, you can use the variations function on an image to create different outcomes by adding more descriptive elements or changing the aspect ratio.

  • What should you do if you want to include an image URL in your prompt without using Discord?

    -You can use an image hosting service like postimage.cc to upload your image, get a direct link, and then paste that link into your prompt in Midjourney.

  • How does the 'variations' function differ from the 'blend' function in Midjourney?

    -The 'variations' function allows you to make slight alterations to a single image to create different outcomes, while the 'blend' function combines multiple images into one prompt, offering a more complex mix of styles and elements.

Outlines

00:00

🌐 Image Prompting in Midjourney: Basics and Advanced Techniques

This paragraph explains the fundamental and advanced methods of using images to generate prompts in Midjourney. It starts with the basic process of obtaining an image URL, either from a website or by uploading an image to platforms like Discord or postimage.cc. The speaker then demonstrates how to adjust the 'Image Weight' (IW) to control the influence of the image on the generated output. The paragraph also covers the use of 'remix mode' for re-rolling prompts with altered image weights, the process of upscaling images, and the concept of 'chain prompting' where AI-generated images are used to create subsequent prompts. Additionally, it introduces the 'blend' function, which allows for combining multiple images into one prompt, and a 'describe' function that generates descriptive prompts based on an uploaded image.

05:02

🔄 Exploring Variations and Combining Describe with Image Prompts

The second paragraph delves into using 'variations' on an image to create different outcomes while keeping the core theme intact. It discusses the ability to change the aspect ratio and other attributes of an image to influence the generated art. The paragraph also highlights the 'describe' function, which provides descriptive prompts based on an uploaded image, allowing users to edit and refine these prompts for more control over the final output. Furthermore, it demonstrates how to combine the 'describe' function with an image prompt to achieve unique and varied results, emphasizing the creative potential of blending text and image-based prompts in Midjourney.

Mindmap

Keywords

💡Image Prompting

Image prompting refers to the process of using an existing image as a reference to guide the creation of a new image by an AI, such as Midjourney. In the video, the user demonstrates how to input an image URL into the prompt to generate a new image inspired by the provided image. This technique is central to the video's theme of exploring different methods to enhance image generation using AI.

💡URL

A URL, or Uniform Resource Locator, is a reference to a web resource that specifies its location on a computer network and a mechanism for retrieving it. In the context of the video, the user explains how to obtain an image URL from various platforms like Discord or using services like postimage.cc, which is essential for using images in the AI image generation process.

💡Discord

Discord is a popular communication platform that allows users to communicate via text, voice, and video. The script mentions using Discord to upload and share images by copying their URL, which can then be used as prompts for image generation in Midjourney. It serves as an example of a platform where users can source images for AI prompting.

💡Image Weight

Image Weight is a parameter in Midjourney that allows users to control the influence an image has on the generated output. The video demonstrates using the '--IW' command followed by a number (up to 2 in the mentioned version) to increase the image's impact on the AI's output. This is a key concept as it directly affects how closely the generated image resembles the source image.

💡Remix Mode

Remix mode is a feature that allows users to re-roll or modify the generated image with different parameters without starting the process from scratch. In the video, the user turns on remix mode to adjust the image weight and add additional prompts to the image generation process, showcasing its utility in fine-tuning the AI's output.

💡Chain Prompting

Chain prompting is a technique where an AI-generated image is used as a prompt to generate another image, which can then be used in another prompt, and so on. This iterative process allows for the continuous evolution of the art. The video illustrates this concept by showing how one can use an upscaled image as a prompt and then generate new images from that, creating a chain of image generation.

💡Blend Function

The blend function is an advanced feature that allows users to combine multiple images into one prompt, creating a blended influence on the AI's image generation. In the video, the user demonstrates how to use the 'slash blend' command to upload and blend five images of Megatron, showcasing how this function can amalgamate different visual elements to generate a unique output.

💡Variations

Variations refer to the process of generating multiple slightly different versions of an image based on a single prompt or image. The video script mentions using the variations command to create different interpretations of the same theme, allowing the user to explore diverse outcomes from a single source of inspiration.

💡Describe Function

The describe function is a tool that provides descriptive prompts based on an uploaded image. It helps users to generate prompts that capture the essence of the image without directly referencing it. In the video, the user uses the 'slash describe' command with an image of Megatron to receive four descriptive prompts, which can then be used to guide the AI in creating new images.

💡AI Art Generation

AI Art Generation is the process of creating visual art through artificial intelligence, often using prompts to guide the style and content of the generated artwork. The video's main theme revolves around AI art generation, specifically using Midjourney to create images inspired by user-provided prompts, images, or descriptive text.

💡Cross-Referencing Images

Cross-referencing images involves using multiple images as prompts to influence the AI's output, resulting in a blend of styles or elements from each image. The script describes how the user can input two image URLs to create a cross between the two, leading to a unique style of result by combining their visual elements.

Highlights

Using a URL of an image as a prompt in Midjourney can inspire a new generation of images.

Obtaining an image URL for use in Midjourney can be done by uploading an image to a platform like Discord or using a service like postimage.cc.

The Image Weight parameter (-IW) allows users to control the influence of the image on the generated prompt, with a maximum value of 2 in version 5.

Remix mode enables re-rolling with different image weights to reference the image less or more in the generated image.

Chain prompting is a technique where an AI-generated image is used as a prompt to generate another image, which can then be used in further prompts.

Blending images can create a mix of styles and elements from two or more images to produce a unique result.

The 'blend' function allows for combining multiple images into one prompt without the need for an additional text prompt.

Descriptive prompts generated by the 'describe' function provide four different descriptive prompts based on an uploaded image.

Variations can be made on the results of the 'describe' function to further refine and customize the generated images.

Combining the 'describe' function with an image prompt can lead to unique and creative outcomes.

The ability to copy and paste URLs of images directly into the Midjourney prompt is a straightforward way to use specific images as inspiration.

Upscaling a favorite image and using it as a prompt allows for detailed referencing down to the pose in the generated image.

Different styles of results can be achieved by adding text prompts to the end of an image URL prompt.

The 'slash blend' command allows users to upload and blend multiple images together for a more complex and layered result.

Adding descriptive text after blending images with the 'slash blend' command can further influence the final image.

Using the 'variations' command with remix mode activated allows for additional descriptive inputs to alter the generated image.

The 'describe' function can be used independently or in conjunction with image prompts to generate a diverse range of image prompts.

Image prompting in Midjourney offers a variety of methods to achieve unique and personalized AI-generated images.