Table of Contents HIDE
In this brief guide, I will teach you how to create stunning AI art from any of your game screenshots using the free service Playground AI. There, you will be able to create an unlimited number of pictures per day (1000 per day without delay, and thereafter with a delay of about 30 seconds). Let me explain the basics of editing your screenshots in Playground AI.
Let’s assume that you have a game screenshot that you want to change. For example, I have this girl from Black Desert.
Crop the screenshot to the resolution supported by the free version of Playground AI, such as 768×512. If the size of the original and generated picture differs, you may encounter problems with proportions.
However, if you do not specify “Image Dimension” explicitly, you will receive images in very low resolution.
Image to Image
Now you should upload your screenshot to the “Image to Image” field and specify “Image Strength“. AI will be “inspired” by your image when generating new ones, and “Image Strength” will determine how much the AI arts will resemble the original screenshot. I recommend setting the value between 40 to 80, although situations may vary.
Prompt for Redrawing the Screenshot
Now let’s write a prompt, in which we will tell AI exactly what we want to see. I prefer to conditionally divide the prompt into 3 parts: image parameters (quality, resolution, etc.), character description, and background description.
You can see what prompts people use in the Community Feed or rewrite mine if you want. In the prompt, we let AI understand that we want to see a beautiful elf archeress standing against a backdrop of a flower field.
Next, you need to enable the “Exclude from Image” option and describe what we do NOT want to see in the picture. For example, incorrect anatomy, naked people, poorly drawn faces, extra limbs, and so on. This is more relevant when generating images from scratch, but it will also be useful when editing screenshots.
Playground AI allows you to apply various filters for image generation. Depending on the theme of the picture, different filters yield different results.
Therefore, I recommend setting the number of images generated in the right panel to 1-2 pieces and checking as many filters as possible.
If you generally like what you see, you can continue generating with the given filter by changing “Prompt Guidance” (recommended 8-12), “Image Strength“, and the prompt itself until AI creates the perfect art for you.
Upscaling the Image
When you generate something beautiful, be sure to enlarge it and save it to your computer. To do this, click “Actions” on the AI-created art and select “Upscale x4“.
Now you can download your beautiful art in high resolution and make it, for example, your desktop wallpaper.
Take a look at the high-resolution result. You can create such arts in just a few seconds if you practice and study the Stable Diffusion prompting.
Tips for Beginners
- The larger the plan and the higher the resolution of the image being generated, the better the result will be. Avoid small details.
- Do not set too high a “Prompt Guidance“. Give the AI more freedom, otherwise, you will often get distorted faces and limbs.
- Do not get stuck on one image if you can’t get a beautiful picture for a long time. Try changing the filter, settings, prompt, or even the source screenshot.
If you have additional questions, you can ask them in the comments. I will try to help to the best of my knowledge and skills 😎.