Remove Objects from Images

It takes a Reference Image and a Target Object description, then creates a Removed Image. The target can be identified by its identity, its color, or both, with the intended result of preserving the rest of the image.

Prepare one image for the Reference Image input and a text description for Target Object. The saved example names “orange chair,” but you can describe another target when using the recipe in Fuser.

Yes. Reference Image and Target Object are exposed inputs that you can edit in Fuser. Replace the image, change the target description, or update both before creating a new Removed Image.

The saved default connects a reference image with the text “orange chair” and an image output labeled Removed Image. It is a saved example rather than a live run, so a new result uses the inputs you provide.

Yes. The recipe supports identifying an object by color as well as by identity, or using both together. A description such as a color-based target can help define what should be removed from the reference image.

It is useful for photographers, artists, and product or industrial teams working with images that contain an unwanted object. The output is intended for image cleanup and alternate visual versions within those areas.

The recipe produces one image output labeled Removed Image. It does not expose separate outputs or additional editing controls in the saved recipe, so the result is centered on removing the target from the supplied reference image.