All terms
Multimodal
Image Captioning
Generating a written description of what appears in an image.
Definition
Image captioning is the task of producing a written description of a given image. A model typically pairs an image reader that picks out what is in the picture with a text generator that writes the caption, trained on large collections of images paired with text. It serves as a foundation for richer image-and-text tasks, and caption quality is often scored by automatic measures that compare the result against human-written captions.