Imagen on Vertex AI lets you edit Imagen-generated or existing images. You can specify part of the image to modify in addition to a text description of the updates (mask-base editing), or you can use a text prompt to update the entire image (mask-free editing).
Types of available editing tools
The following are types of editing that Imagen supports:
Mask-based editing - Lets you specify a targeted area to apply edits to. This method of editing is good for edits that apply to only parts of an image. Mask-based editing is useful for the following types of image edits:
Mask-free editing - Lets you edit an image without a mask (specific area). This method of editing is good for edits that belong to the entire image, or when the location of the edits is not as important for your use case. For location-specific editing, use mask-based editing.
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
Both input data and output content are checked for offensive material when you send an image editing request to Imagen. This means offensive input data from a text prompt or from a base image you provide can be blocked. Similarly, offensive output images might also be blocked, affecting the number of generated images you get in a response.
For more information about safety filtering and blocked content handling, see Responsible AI and usage guidelines for Imagen.
Model versions
There are multiple image generation models that you can use. Each model may have limits that are specific to that model. For more information, see Imagen models.