Edit images overview

This page provides an overview of the image editing features in Imagen on Vertex AI. You will learn about the following:

• Types of available editing tools: Learn about the differences between mask-based and mask-free editing.
• Locations: Learn which regions support image editing models.
• Safety filtering: Understand how safety filters are applied to your input and output images.
• Model versions: Review the available image models and their specific limits.

Types of available editing tools

You can edit Imagen-generated or existing images using one of two methods: mask-based editing or mask-free editing.

• Mask-based editing: With this method, you specify a targeted area (a mask) to edit, while the rest of the image remains unchanged. This method is ideal for ___location-specific edits, such as the following:

  • Insert objects into an image
  • Remove objects from an image
  • Expand the content of an image
  • Replace the background of an image

• Mask-free editing: This method lets you edit an entire image using only a text prompt, without defining a specific area. It is best for global edits that affect the entire image or when the precise ___location of the change is not important. For more information, see Edit images using text prompts.

Locations

A ___location is a region you can specify in a request to control where data is stored at rest. For a list of available regions, see Generative AI on Vertex AI locations.

Safety filtering

When you send an image editing request, Imagen checks both your input data and the generated output for offensive material. As a result, the service might block offensive input from a text prompt or a base image. The service might also block offensive output images, which affects the number of generated images you receive in a response.

For more information about safety filtering and blocked content handling, see Responsible AI and usage guidelines for Imagen.

Model versions

You can use several image generation models. Each model might have specific limits. For more information, see Imagen models.

What's next

• Insert objects into an image
• Remove objects from an image
• Expand the content of an image

• Edit product images

• Edit images using text prompts.