HOW TO MAKE WRINKLE ASSETS WITH AI

ether
29 Jan 202407:56
EducationalLearning
32 Likes 10 Comments

TLDRThe video demonstrates how to use AI image generation to create realistic wrinkled clothing assets for games or 3D projects. It introduces Leonard AI, a free AI art generator, and shows how to upload reference photos, customize prompts with details like resolution and realism, and generate multiple wrinkled clothing images. It then shows how to incorporate these AI-generated images into clothing asset templates in a program like Paint.net, enhancing them with effects like clarity to make them look more convincing. The video aims to provide an efficient way to create quality wrinkled clothing assets without having to manually create or find photo references.

Takeaways
  • 😊 The video demonstrates how to use AI to generate wrinkled cloth assets for games/CG
  • πŸ‘¨β€πŸ’» The AI software used is called Lonardo AI, which provides 150 free credits per day
  • πŸ“Έ Image guidance is used to provide the AI a reference photo to generate assets from
  • πŸ–ΌοΈ Multiple generated images are edited in Paint.net to create asset template textures
  • 🌟 Highlight/shadow tweaks make the final assets look more realistic
  • πŸ‘• The process works for shirts too without needing an image reference
  • 🎨 A prompt describes the asset to generate - specifics like resolution help
  • πŸš€ With practice, custom AI software could be made for this purpose
  • ⏱️ Using AI is much faster than manually finding & editing texture images
  • 🀝 The video aims to introduce viewers to using AI for game design tasks
Q & A
  • What AI software is used in the video?

    -The video uses an AI application called Lonardo AI to generate wrinkled clothing images.

  • Is Lonardo AI free to use?

    -Lonardo AI provides 150 free points per day which is enough to generate the needed AI images for clothing assets.

  • Why is image guidance needed when generating jeans images?

    -Image guidance with a reference photo is needed when generating jeans images because jeans have a complex and specific wrinkle pattern that the AI needs guidance on.

  • What image dimensions are used?

    -The video sets the image dimensions to 768x768 pixels when generating the AI images.

  • What is the benefit of using AI for wrinkle images?

    -Using AI for wrinkle images allows you to get more variant images from just one reference photo rather than manually searching for many different reference images.

  • Is image guidance needed for generating t-shirt images?

    -Image guidance is not needed for generating basic t-shirt images since they have a more simple wrinkle pattern.

  • What prompt is suggested for t-shirts?

    -A suggested prompt is 'white t-shirt on white backdrop with wrinkles' along with settings like 4K, high res, and realistic.

  • What post processing improves the asset?

    -Increasing the clarity to 100% in post processing adds detail and makes the AI generated asset look more realistic.

  • Why use color t-shirts instead of white?

    -Using a colored t-shirt instead of white generates better results in the AI since details show up better on colors.

  • What is the overall goal of using this AI technique?

    -The goal is to streamline the process of creating realistic wrinkled clothing assets by leveraging AI generation.

Outlines
00:00
πŸ˜„ Explaining how to use AI to generate wrinkled clothing assets

The narrator introduces using an AI application called Leonard AI to generate wrinkled clothing assets rather than creating them manually. He shares his screen to demonstrate finding reference images on Depop, preparing them in Paint.net, and configuring the settings in Leonard AI to generate wrinkled clothing. He shows examples of generated jeans and t-shirt assets, advising to adjust clarity to make them look more realistic.

05:00
πŸ‘• Tips for making the AI-generated assets look realistic

The narrator explains why Leonard AI is better than other methods to make wrinkled assets since it allows more variance from a single image. He recommends viewers learn how to make clothing assets if they don't know already. To make the AI-generated assets look more realistic, he strongly advises adjusting the highlights and clarity slider in editing software up to 100% to add grain and noise like the reference picture.

Mindmap
Keywords
πŸ’‘Wrinkle assets
Wrinkle assets refer to graphic design assets or images that have wrinkles, folds, or creases, especially for depicting clothing. They are used to add realism and details when designing clothes and fabric images. The video describes how to generate realistic wrinkle assets using AI instead of making them entirely manually. The assets are applied to shirt and jeans templates in the video to demonstrate the process.
πŸ’‘AI
AI stands for artificial intelligence. In the video, it refers to an AI system called Lonardo AI that can generate realistic wrinkled clothing images when provided with appropriate prompts and source images. The AI capabilities are leveraged to simplify and expedite the wrinkle asset creation process compared to manual techniques.
πŸ’‘Image guidance
Image guidance is a feature in Lonardo AI that allows providing an example source image to guide the AI in generating new images. For wrinkle assets, an image of wrinkled jeans or shirt is uploaded to steer the AI to produce similar wrinkled results.
πŸ’‘Prompts
Prompts refer to text descriptions provided to the AI to specify what type of image to generate. Details like clothing color, style, resolution etc. are provided in prompts to produce relevant wrinkled assets.
πŸ’‘Paint.net
Paint.net is a basic image editing software used in the video. It is used to crop and extract sections of reference images to prepare image guidance for the AI generator.
πŸ’‘Variants
Variants refer to generating multiple different versions of wrinkled shirt/jeans images from the AI instead of just one. This adds more uniqueness and options for using them as assets.
πŸ’‘Highlight & Shadow
These are effects in Paint.net to adjust contrast and lighting. Boosting clarity and shadow/highlight separation is done to make the AI generated assets look more realistic like true clothing.
πŸ’‘Assets
Assets are reusable graphical elements like images, textures, 3D models etc. Wrinkled shirt and jeans images are saved as assets to easily apply on clothing designs instead of drawing wrinkles manually each time.
πŸ’‘Reference pictures
Reference pictures refer to source images, like wrinkled jeans photos from a clothing website, that provide examples for the AI to generate new wrinkle asset variants from.
πŸ’‘Realistic
Realistic in the context of the video means wrinkle assets that closely resemble the natural wrinkles and folds in real clothing. The goal is to achieve realism through AI generation and post-processing.
Highlights

Using AI to generate wrinkled clothing asset images

Lonardo AI is a free AI image generator with daily credits

Upload a reference image to guide the AI generation

Use a prompt to specify details like resolution and realism

Generate multiple variants from a single reference image

AI generated images may look unrealistic or distorted

Post-process the images to increase realism

Works for wrinkled shirts without a reference image

Avoid white backgrounds for best results

Paste AI images into templates to create assets

Increase clarity to 100% to add grain and noise

Makes the asset look more realistic

Easier than searching for multiple reference images

Saves time compared to manual methods

Opens up new creative possibilities with AI

Transcripts
Rate This

5.0 / 5 (0 votes)

Thanks for rating: