Leonardo AI New Feature: Consistent Characters with Text to Image Art Generators - AI Tutorial
Education
Introduction
Hi everyone, my name is Iman, and this is the Iman Global YouTube channel. Welcome back to another AI tutorial. In this tutorial, I'm going to show you a new feature of Leonardo AI for creating consistent characters. Please support my YouTube channel—don't forget to like and subscribe! Let's get started.
Are you struggling to create consistent characters with AI image generators? Meet the new character reference feature in Leonardo AI. This innovative tool lets you design a base character that stays consistent in different poses, outfits, and backgrounds. Say goodbye to morphing faces and hello to cohesive stories and artwork. Ready to elevate your AI image generation skills? Join me in this tutorial and explore the art of consistent character creation with Leonardo AI.
Getting Started
We are visiting the Leonardo website using the launch app button located in the top right corner. I create a user account, and after creating the account, I log in. We are now on the Leonardo AI dashboard. From the left-hand menu, I click on the image generation option. As you can see, the interface for the image generation section has changed and been updated. However, by activating the legacy mode button in the top right corner, you can switch back to the previous classic image generation interface. I prefer the new interface.
Step 1: Creating a Character
In the first step, we need to create a character to place in different poses, outfits, and locations. I have prepared a text prompt for creating this character. In the new interface, there is a text box for entering the prompt at the top of the site. I paste the prompt into the text box.
From the left section in the preset drop-down list, I select the Portrait Perfect option. A bit further down from the preset style drop-down list, I select Portrait Bokeh. In the contrast drop-down list, I choose Medium. In the generation mode section, you can toggle between fast generations and slower quality generations. Quality is formerly known as Alchemy V2. I select the Quality option.
I set the image dimensions to horizontal with a 16:9 aspect ratio. A bit further down for the image size, I choose the Large option, which corresponds to 2384 by 1344 pixels. For the number of images, which indicates how many images are generated at each time, I set it to four. With the private mode option, we can hide our generation from the community feed.
I then go to the advanced settings section. From the model drop-down list, I set it to Leonardo Kino XL. Enabling the Photo Real option results in a high-quality finishing pipeline that produces more photorealistic outputs. By enabling the negative prompt option, a new section is added to the prompt box where you can enter what you do not want to see in the image. For example, to avoid generating colored hair, I enter a list of hair colors here.
Now I click the Generate button. After a short while, the images are generated based on the prompt and settings we configured. I review these images. Note that there is a download button below each frame to download the images for free. I have chosen the image you see as my main model.
Step 2: Placing the Character in Different Poses, Outfits, and Locations
Now that my main model is created, in the second step, I need to place it in different poses, outfits, and locations. To do this, I click on the icon located in the top left corner of the prompt box. A list opens, and from this list, I select the Character Reference option, which is used to create a consistent character by providing a reference image. A new window for selecting the character reference opens. Here, I can set the character generated in the previous step.
I click on the Your Generation tab. In this section, I select one of the images we want to keep consistent, which we created in the previous step. I click on the Confirm button. The thumbnail of this image appears in the bottom left corner of the prompt box. I click on the downward arrow on this thumbnail. A list opens where we can choose between Low, Medium, and High to determine the impact of the reference image on subsequent generations. I select High.
Now, I delete the previous prompt and define a new pose, outfit, location, or situation for the character to be generated in. With the same settings as before, I click on the Generate button. After a short while, new images are generated. These images feature the same model with the same body anatomy and appearance, or the same character reference, in a new situation with a new outfit, in a new location, and in a new pose.
Reviewing New Images
This way, with the new Leonardo AI feature, we achieve a consistent character. As I review the images, you can see that there is a strong resemblance between the main model image and the new images, and the consistent character is maintained.
Now I want to place the main character or character reference in another new situation with a new outfit, new location, and new pose. I only need to keep the same previous settings and just change the text prompt. I delete the previous prompt and enter a new prompt in the text box. Again, I click on the Generate button. After a short while, new images are generated. These new images feature the same model with the same body anatomy and appearance, or the same character reference, in a new situation with a new outfit, in a new location, and in a new pose.
I review the images. You can see that there is a strong resemblance between the main model image and the new images, and a consistent character is maintained.
Conclusion
I hope this tutorial has helped you solve the inconsistent character issue and that you can now achieve consistent character generation with Leonardo AI. Thank you for watching this tutorial. If you learned something new, please like and share it with your friends. If you have any questions about AI, please don't hesitate to ask. I'm always happy to help. There are many more AI tutorials on the way, so subscribe and hit the notification bell to stay updated. The future is AI. Embrace it. See you in the next tutorial!
Keywords
- Leonardo AI
- Consistent Characters
- AI Image Generation
- Character Reference Feature
- AI Tutorial
FAQ
Q: What is the character reference feature in Leonardo AI? A: The character reference feature in Leonardo AI allows you to create a base character that remains consistent in various poses, outfits, and backgrounds.
Q: How do I switch between the new and classic image generation interface? A: You can switch between the new and classic image generation interface by activating the legacy mode button in the top right corner of the image generation section.
Q: What settings should I use to achieve consistent characters in Leonardo AI? A: To achieve consistent characters, create a base character using a specific prompt and then use the character reference feature to maintain consistency across different poses, outfits, and backgrounds.
Q: Can I hide my generated images from the community feed? A: Yes, you can hide your generated images from the community feed by enabling the private mode option.
Q: What is the photo real option in advanced settings? A: The photo real option in advanced settings results in a high-quality finishing pipeline that produces more photorealistic outputs.