All terms
Multimodal
Image to Image
Transforming one image into another, optionally guided by a text prompt.
Definition
Image to image is the task of transforming a source image into a new one rather than generating from scratch, often guided by a text prompt or other input. Examples include applying the look of one image to another, turning a sketch into a photo, sharpening or enlarging a picture, and letting an image-making AI begin from an existing image instead of from random noise. It underpins many editing and refinement workflows.