How to change clothes with AI.

Sebastian Kamph
21 Jan 202414:44

TLDRThe video tutorial demonstrates how to use stable diffusion and in-painting techniques to change the clothes in an image. The presenter guides viewers through the process of altering an image of a woman in a white crop top and blue jeans, showing how to modify the clothing to different styles and materials such as leather pants or a silk top. Tips are provided for dealing with imperfections, and the use of in-painting models is discussed. The video also covers adding new clothing items to an image using the in-painting sketch feature. The presenter emphasizes the importance of experimenting with different settings, such as denoising strength and masked content, to achieve the desired results. The tutorial concludes with a reminder to check out the presenter's Patreon for pre-loaded in-painting styles, offering an easier way to enhance the AI's capabilities.


  • 🎨 **Image Transformation**: The video demonstrates how to change an image using AI, specifically through stable diffusion, to alter clothing styles.
  • 🖌️ **Inpainting Tool**: Utilizes the inpainting tool for detailed editing, allowing for the modification of specific parts of an image, like changing clothing.
  • 📱 **Zoom Feature**: Pressing 's' provides a zoomed-in view for precise painting, which is crucial for detailed work.
  • 👖 **Changing Clothes**: The process involves changing a pair of blue jeans to leather pants and a white crop top to a green silk top, showcasing the versatility of the technique.
  • 🔍 **Masking Technique**: The importance of properly masking the area to be changed is emphasized to ensure the AI focuses on the correct part of the image.
  • 🌈 **Color and Material Change**: The AI can alter both the color and material of clothing items within an image, offering a wide range of creative possibilities.
  • 📈 **Denoising Strength**: Adjusting the denoising strength can help in achieving the desired result, with higher values introducing more changes to the image.
  • ✅ **Success Rate**: The process involves trial and error, as the AI's generative nature can produce varying results, requiring multiple attempts.
  • 🧩 **Iterative Process**: To continue editing the same image, it must be dragged from the right to the left within the tool, indicating an iterative approach.
  • 🧵 **Adding Clothing**: The video also covers how to add new clothing items to an image using the inpainting sketch feature.
  • 🔗 **Styles and Patreon**: The presenter mentions that they have preloaded painting clothing styles available through their Patreon, which can enhance the editing process.

Q & A

  • What is the main topic of the video?

    -The main topic of the video is how to change clothes in an image using AI, specifically with stable diffusion and inpainting techniques.

  • What is stable diffusion?

    -Stable diffusion is a type of AI model used for generating images from textual descriptions, which can also be used for image-to-image translation and inpainting.

  • What does the speaker mean by 'inpainting' in the context of the video?

    -In the context of the video, 'inpainting' refers to the process of editing an image to fill in or alter certain parts of it, such as changing the clothes of a person in the image.

  • Why is an inpainting model suggested for use in the video?

    -An inpainting model is suggested because it can provide improved results for specific use cases where the AI needs to fill in or modify parts of an image, such as changing clothing textures and styles.

  • What is the purpose of the 's' key in the inpainting process as described in the video?

    -The 's' key provides a zoomed-in view of the image, allowing for more precise drawing and editing within the inpainting process.

  • How can one change the style of clothing in the image?

    -One can change the style of clothing by using preloaded inpainting cloth styles or by manually entering styles and finding suitable negatives for the desired material in the prompt.

  • What is the significance of the 'masked content' setting in the inpainting process?

    -The 'masked content' setting determines what the AI uses as a reference for generating the new content. It can be set to 'original' to retain the existing style, 'fill' for a solid color, or 'latent noise' to introduce new patterns and textures.

  • Why might one choose to generate the image at a lower resolution for the pants but blend it into the full image?

    -Generating the image at a lower resolution for the pants can help blend it better with the rest of the image, even though it may result in less detail for that specific area.

  • How does the 'D noising strength' setting affect the outcome of the inpainting process?

    -The 'D noising strength' setting controls the level of detail and randomness in the generated content. A higher value introduces more changes and can create a completely new style, while a lower value retains more of the original structure.

  • What is the 'in painting sketch' feature used for in the video?

    -The 'in painting sketch' feature is used for adding new elements to the image that were not originally present, such as drawing a hat and then using the AI to generate a detailed version of it.

  • What is the importance of testing different settings and iterations when using AI for image editing?

    -Testing different settings and iterations is important because the outcome can be somewhat random, and it allows for fine-tuning the process to achieve the desired result, balancing between detail retention and introducing new elements.



🎨 Introduction to Image Transformation with Stable Diffusion

The video begins with an introduction to transforming an image using stable diffusion and in-painting techniques. The presenter demonstrates how to change the clothing of a character in an image and offers tips for troubleshooting. They use an image of a woman in a white crop top and blue jeans as a starting point and guide viewers through the process of changing her clothes to different styles and materials, such as leather, using preloaded styles available through Patreon.


👖 Modifying Pants Material and Color

The presenter focuses on changing the material and color of the pants in the image. They explain the process of in-painting, selecting the right settings, and adjusting the denoising strength for better results. The video shows how to achieve a natural look with black leather pants and discusses the importance of blending the new material with the rest of the image. It also touches on the iterative process of image generation and the need for trial and error.


🧥 Changing the Crop Top and Adding Clothing

The video continues with changing the material and color of the woman's crop top to silk and other materials, and then to a red latex glossy top. The presenter discusses the use of latent noise for introducing new data and achieving different colors or patterns. They also cover the process of adding new clothing items to the image, such as a red leather hat, using the in-painting sketch feature. The video concludes with a reminder that generative AI involves a degree of randomness, and that achieving the desired result may require several attempts.



💡Stable Diffusion

Stable Diffusion is a term referring to a type of artificial intelligence model used for generating images from textual descriptions. In the context of the video, it serves as the foundation for changing the appearance of clothing in an image. The script mentions using Stable Diffusion to generate an initial image of a woman with specific clothing, which will then be altered using the in-painting technique.


In-painting is a technique used in image processing to fill in or smooth out selected parts of an image. In the video, the presenter uses in-painting to modify the clothing of a character in an image by first masking the area to be changed and then applying a new style or material to that area.


Masking in the context of image editing refers to the process of selecting a specific area of an image to isolate it for modification. The script describes using a tool to draw a mask around the clothing items that need to be changed, which is a crucial step before applying new styles or materials.

💡Denoising Strength

Denoising Strength is a parameter in image processing models that controls the level of detail and noise introduced into the generated image. In the video, adjusting the denoising strength is discussed as a way to control the extent of changes made to the clothing in the image, with higher values leading to more significant alterations.

💡Latent Noise

Latent Noise is a concept in AI image generation that involves introducing random variations into the image to create new patterns or textures. The script mentions using latent noise when changing the material of clothing to generate a completely new look, as opposed to relying on the original image data.


Styles in the context of the video refer to the different types of clothing materials or patterns that can be applied to the masked area of an image. The presenter discusses using preloaded in-painting cloth styles such as cotton, leather, silk, and velvet to change the appearance of the clothing in the image.

💡Image-to-Image Tab

The Image-to-Image Tab is a feature within the AI image generation tool that allows users to modify existing images rather than starting from scratch. The video script describes navigating to this tab as part of the process to change the clothing in an image.


Resolution in digital imaging refers to the detail an image will hold. It is often measured in pixels, such as 768 x 768 as mentioned in the script. When generating a new image or modifying a part of an image, the resolution can affect the level of detail and clarity in the final output.


Patreon is a membership platform that allows creators to receive financial support from their audience in exchange for exclusive content or benefits. In the video, the presenter mentions Patreon as a source where viewers can find the preloaded in-painting cloth styles used in the demonstration.

💡Auto-Detect Ruler

The Auto-Detect Ruler is a tool within the image editing software that automatically determines the correct resolution for an image based on its content. The script describes using this tool to correct an issue where the resolution was set incorrectly, ensuring the generated image matches the intended quality.

💡In-Painting Sketch

In-Painting Sketch is a feature within the AI image generation tool that allows users to add new elements to an image that were not originally present. The video demonstrates using this feature to draw and add a hat to the character in the image, showcasing the tool's ability to generate new content based on simple sketches.


Learn how to change clothes in an image using AI and stable diffusion techniques.

The process is super simple and can be done by anyone after watching the video.

Tips and tricks are provided for when things don't work out as expected.

A starting image is used, generated through stable diffusion, featuring a woman in a white crop top and blue jeans.

Different styles of clothing can be applied, such as leather, silk, and wool, in various formats.

Inpainting models can be used for special cases where they perform better.

The 's' key allows for a zoomed-in view for more precise painting.

Painted areas can be masked or unmasked depending on the desired outcome.

Changing the masked content to latent noise can help with color changes.

Denoising strength can be adjusted for better detail retention or more significant changes.

The process involves trial and error, with results varying between iterations.

If an error occurs, ensure a mask has been properly drawn around the area to be changed.

Adding clothing that wasn't originally in the image can be done using the inpaint sketch feature.

The material of existing clothing can be changed without altering the color.

Velvet and latex are examples of materials that can be applied to change the look of clothing.

The final result can significantly change a person's appearance in an image through AI.

The video provides a powerful workflow for altering clothing in images using generative AI.

Patrons can access preloaded inpaint styles for an easier experience.