When generating images, even with great text-to-image prompts, it's possible that some small details require adjustments due to the inherent randomness of image generation models. To tweak your generated images, you can use the Canvas Editor.
1. Navigate to the Canvas Editor
After generation, you can hover over an image to reveal the Canvas Editor represented by a square icon. Click on the square Canvas Editor icon to bring that image into the Canvas Editor.
2. Frame the area you would like to edit
To frame the area you would like to make edits within, click on the Select icon on the left-hand menu. Then, drag the selection box to the target area. Please note that any changes made will occur only within the purple frame.
Click on Draw Mask icon on the left-hand menu. Then, draw a mask over the area that you would like to edit.
3. Generating Images in the selected area
First, set the generation settings on the sidebar on the right as follows.
Select Leonardo Diffusion XL as the Finetuned Model, the same model the image was generated from. Leonardo Diffusion XL is a flexible model that delivers stunning outputs even with short prompts.
Select Inpaint / Outpaint as Canvas Mode and disable the Outpaint option. This would effectively enable Inpainting mode.
Set Inpainting Strength to 1. The higher the strength the more the output will diverge from the original image.
Set the Number of Images to 4 so that we can choose among 4 image options later.
Type in your prompt at the prompt bar at the bottom, in this case, we want "a white flower". Click on the Generate button.
The process of generating over an area of an image is called "Inpainting". The AI model will take in the target (masked area) and consider the surrounding context when generating images.
💡 Leonardo's Tip: In typical cases, it would be best to experiment with the Inpainting Strength to find the best results for an image. The ideal amount will vary from image to image.
4. Select and apply the result
Use the arrow keys to toggle across the generated options. Then, click the Accept button to apply the edit.
5. Downloading the Edited Image
Finally, you can download your image by clicking on the Download Artwork button on the left-hand menu.
In this guide, you learned how to use the Canvas Editor to tweak your images by utilizing inpainting. This workflow also works for images not generated within Leonardo.Ai. Simply upload your own image by clicking on the Upload Image on the left-hand menu.
To learn more about the Canvas Editor, check out our in-depth guide on How To Use Canvas Editor.
Aside from editing, Leonardo.Ai also offers other post-processing tools such as removing backgrounds and upscaling. You can learn about these tools in our guide to Refine Images with AI Postprocessing.