Krita AI Trick to Removing Objects With AI diffusion Plugin

In this informative video, the creator explores the use of Krita's AI diffusion plugin to remove objects from images. They introduce a method involving the use of a static image uploaded to their cloud, which viewers can download. The process involves manipulating the AI to recognize and render the static image, effectively replacing the unwanted object with elements from the background. The video also compares the AI diffusion capabilities of Krita with Adobe, highlighting the benefits of using open-source software and the potential for future improvements with new models.

Takeaways

• 🎨 The video discusses a method to remove objects from images using Krita's AI diffusion plugin.
• 🔍 Krita lacks a dedicated AI background removal tool, but the host explores a workaround using AI diffusion.
• 🌟 A 'static diffusion' image is introduced as a tool to aid in manipulating the AI for specific tasks.
• 🖼️ The host shares a downloadable 'static' file to be used with the AI for better image manipulation.
• 📱 Krita's interface is navigated to show where the add-on and static file should be placed.
• 💻 The video highlights the limitations of an older system with an GTX 1070 for high-resolution rendering.
• 🔄 Comparisons are made between Krita and Adobe's AI models, with Krita favored for model selection and live preview.
• 🎨 The process of removing a vehicle from an image is demonstrated, showing the steps and adjustments made.
• 🛠️ Tool options and layer adjustments are crucial for achieving a blended and natural look in the final render.
• 🌐 The host expresses a preference for free, open-source tools like Krita and encourages support for their development.
• 📢 The video concludes with a call to action for viewers to support the creators of such plugins and to share the knowledge.

Q & A

What is the main topic of the video?

The main topic of the video is a tutorial on using the AI diffusion plugin in Krita to remove objects from images.

What issue does the video address with Krita's AI background removal tool?

The video addresses the issue that Krita doesn't have a decent AI background removal tool, and the presenter is exploring ways to manipulate the AI to achieve this.

What is the purpose of the static diffusion file provided by the presenter?

The static diffusion file is provided to help users take advantage of when using the AI in Krita, as it creates a static image that the AI uses to build a unique image every time.

How does the presenter plan to manipulate the AI in Krita to remove an object from the background?

The presenter plans to manipulate the AI by uploading a static diffusion file and using it in combination with Krita's AI diffusion plugin to create a new image without the unwanted object.

What is the advantage of using a lower resolution in the AI diffusion process?

Using a lower resolution in the AI diffusion process can make the renders run faster, especially on older or slower systems, and the presenter mentions that AI upscaling can be used afterward to regain detail without significant loss.

What does the presenter suggest about the selection tool in Adobe's AI system?

The presenter suggests that Adobe's selection tool might work better for certain tasks, as it has a better language model associated with the diffused images, which could provide a better understanding of user inputs like 'remove'.

Why does the presenter prefer Krita's AI diffusion over Adobe's?

The presenter prefers Krita's AI diffusion because it allows for the selection of different models, has a live preview feature that is great for kids and other users, and offers a graphic user interface similar to Adobe or Corel. Additionally, the presenter appreciates that it is free and open source.

How does the presenter suggest using the static diffusion pattern?

The presenter suggests using the static diffusion pattern by uploading it to the user's cloud, downloading it, and placing it in the Krita data folder's patterns, which will then be used in the AI diffusion process to help remove elements from a photo.

What is the presenter's opinion on the future of AI diffusion tools?

The presenter is excited about the future of AI diffusion tools, especially as models evolve and support higher resolutions like 2K and above. They hope for the development of more tools and plugins, including possibly AI removal tools, and encourage support for the creators behind these open source projects.

How can viewers support the creator of the static diffusion plugin?

Viewers can support the creator of the static diffusion plugin by leaving positive feedback on their Reddit posts or other social media platforms, and by using the plugin and sharing their experiences.

Outlines

00:00

🎥 Introduction to AI Manipulation in Stream Tabulous

The video begins with a welcome to Stream Tabulous, where the host expresses an intention to explore AI manipulation techniques, specifically focusing on Cryer, an AI diffusion tool. The discussion includes plans to create videos on how to influence AI to perform certain tasks, such as adding or removing objects within images. The host references a previous video on adding elements like cats to images and plans to address background object removal in this session. The lack of a dedicated AI background removal tool is acknowledged, and the host intends to demonstrate alternative methods to achieve similar results.

05:01

📁 Navigating Cryer and Utilizing Static Diffusion

The host dives into the practical aspects of using Cryer, starting with opening the app and navigating to the relevant data files. The process of installing Cryer and adding Python files is briefly mentioned. The main focus is on the use of a static image in AI diffusion, which serves as a foundation for creating unique images. The host introduces a new static diffusion file that viewers can download from the cloud, emphasizing its role in shaping AI-generated images. The segment also touches on the limitations of the host's system and the impact on rendering capabilities, comparing Cryer's performance to Adobe's AI model.

10:02

🖼️ Manipulating Images and Removing Unwanted Elements

The host demonstrates how to manipulate images in Cryer, particularly focusing on removing an unwanted vehicle from a scene. The process involves adjusting the image resolution to accommodate the host's older system and experimenting with different AI settings to achieve the desired outcome. The host discusses the concept of 'painting' selections in AI, drawing parallels with other AI tools like Stable Diffusion and Adobe's rendering process. The segment also explores the use of static diffusion patterns to influence AI image generation, with the host sharing insights on how these patterns can be used to 'trick' the AI into creating specific images.

15:05

🌳 Fine-Tuning AI Imagery with Nature Elements

The host continues to work on the image, discussing the addition of natural elements like bushes and trees to replace the removed vehicle. The conversation turns to the nuances of AI rendering, including the handling of padding and blending to achieve realistic results. The host shares personal experiences with Adobe's AI and expresses a preference for Cryer due to its open-source nature and the ability to select models that best suit the artwork. The segment also touches on the potential of future AI models, particularly those trained on higher resolutions and better language understanding.

20:07

💡 Experimenting with Reflective Elements and AI Diffusion

In the final part of the video, the host experiments with adding reflective water to the image, comparing the results with Adobe's handling of reflective elements. The host's curiosity leads to a discussion on the variability of AI models and their effectiveness in rendering different scenes. The video concludes with the host's reflections on the potential of Cryer and AI diffusion, advocating for more people to learn about and contribute to the open-source community. The host encourages viewers to download the static pattern file shared on My Cloud and to provide feedback to the creators of the AI tools discussed.

🙌 Encouraging Community Support for Open-Source AI Tools

The host wraps up the video by emphasizing the importance of community support for open-source projects like Cryer. The host expresses gratitude for the developers and contributors to the AI diffusion community, highlighting the value of free tools that can be improved and expanded upon by anyone. The host also mentions the potential for new plugins and tools that could further enhance AI image manipulation capabilities. The video ends with a call to action for viewers to engage with the content, share the video, and support the creators of the AI tools by providing positive feedback and contributions.

Keywords

💡Krita AI

Krita AI refers to the integration of artificial intelligence within the Krita software, which is a free and open-source digital painting application. In the context of the video, Krita AI is used to manipulate images, specifically to remove unwanted objects from a scene. The AI diffusion plugin mentioned is a tool within Krita that leverages AI to generate or alter images based on certain inputs, in this case, to remove a background object.

💡AI diffusion

AI diffusion is a technique that uses artificial intelligence to create or modify images by generating new pixel data based on a set of input parameters. In the video, the AI diffusion plugin is utilized to manipulate the background of an image to remove an object. It works by creating a static image that the AI uses as a reference to build a unique image each time, which can be utilized to replace the unwanted object with a more suitable background.

💡Cloud

In the context of the video, 'cloud' refers to a cloud storage service where the presenter uploads a file that viewers can download. This file is a static diffusion pattern that is used within the Krita AI plugin to aid in the manipulation of images. Cloud storage services are remote servers that allow users to store and access data over the internet, facilitating easy sharing of files across different devices and locations.

💡Patterns

Patterns in the video script refer to the specific types of static images or textures that can be used within the AI diffusion process in Krita. The presenter mentions uploading a static pattern file to the cloud for viewers to download and use within their Krita software. This pattern acts as a base for the AI to generate a new image from, which can help in tasks like removing objects from a scene.

💡Static image

A static image, as used in the video, is a fixed and unchanging visual representation, as opposed to a dynamic or moving image. In the context of AI diffusion, a static image serves as the foundation for the AI to build upon and generate a new, unique image. The static image does not change during the AI's generation process but influences the final output.

💡Resolution

Resolution in the context of the video refers to the pixel dimensions of an image, which determines its level of detail and clarity. Higher resolutions mean more pixels and thus more detail, while lower resolutions result in fewer pixels and less detail. The presenter mentions adjusting the resolution of the image being worked on to accommodate for the limitations of their system's processing power.

💡Adobe

Adobe in the video refers to Adobe Inc., a software company known for its creative and multimedia software suite, including Photoshop and Adobe Illustrator. The presenter compares the capabilities of Adobe's AI tools with those of Krita AI, noting differences in performance, model training, and subscription costs. Adobe's AI features are highlighted as having benefits such as better language models and the ability to handle higher resolutions, but the presenter also expresses a preference for Krita due to its open-source nature and cost-free availability.

💡Stable diffusion

Stable diffusion is a term used in the context of AI-generated images, referring to a model that is trained on a specific set of data to produce stable and consistent image outputs. The video mentions stable diffusion in relation to the AI models used within Krita and other software, suggesting that these models are trained on specific image sizes, such as 512x512 or 124x124 pixels, to create images with varying levels of detail.

💡Open source

Open source refers to a type of software licensing where the source code is made publicly available, allowing anyone to view, use, modify, and distribute the software freely. In the video, the presenter expresses gratitude towards the open-source community for making tools like Krita AI accessible to everyone without cost. The open-source nature of Krita and its plugins is contrasted with the subscription model of Adobe's software.

💡Reflective water

Reflective water in the context of the video refers to the visual effect of water that mirrors or reflects the surroundings, such as objects, light, and colors. The presenter is curious about whether the AI diffusion plugin can replicate this effect as well as Adobe's tools, which were noted for their ability to create realistic reflective water in images. This relates to the overall theme of exploring the capabilities and limitations of Krita AI in comparison to other software.

💡GitHub

GitHub is a web-based platform that provides version control and collaboration features for software development projects using Git. It is where developers can store, manage, and collaborate on their codebases. In the video, the AI diffusion plugin for Krita is mentioned to be available on GitHub, indicating that the plugin's source code is open and accessible for the community to view, modify, and enhance.

Highlights

The video discusses a method to manipulate AI diffusion in Krita to remove unwanted objects from an image.

The use of a static diffusion pattern is introduced to aid in the removal process.

A file is provided for download to assist with the AI manipulation process.

The video explains how the AI creates a static image to build a unique image each time.

The importance of selecting the right model for the task at hand is emphasized.

A step-by-step guide on how to install and use the AI diffusion tool in Krita is provided.

The video demonstrates how to adjust the resolution to improve rendering speed on older systems.

The process of using the static diffusion pattern to replace an unwanted vehicle in an image is shown.

The video suggests that the AI may read the static pattern as a new render to build an image.

Adjustments to padding and blending settings can improve the final render's integration with the surrounding image.

Different models may produce varying results, and it's recommended to choose models suited to the artwork's style.

The video compares the AI diffusion tool in Krita with Adobe's AI model, highlighting the benefits of each.

The potential for future sdxl models to support higher resolutions like 2K is mentioned.

The benefits of using open-source tools like Krita for cost savings and flexibility are discussed.

The video explores the possibility of reflective water rendering using the AI diffusion tool.

The importance of providing feedback and support to open-source developers for continued improvements is emphasized.

The video concludes with a call to action for viewers to share their experiences and provide positive feedback.