Which type of caption-image relationship relies on the image to give meaning to the caption?

Which type of caption-image relationship relies on the image to give meaning to the caption?

• A. Caption-driven
• B. Context-driven
• C. Universal
• D. Picture-specific

Answer: (D)

Grounding between caption and image means the visual content is essential to understand the caption. In a picture-specific relationship, the image provides the concrete details that give the caption its full meaning—without seeing the image, the caption would be vague or ambiguous. The image supplies who, what, where, and how, so the caption only makes sense when paired with that visual context. For example, a caption like “a person with a bright umbrella” becomes clear and specific when you see the person, the umbrella’s color, and the scene. The meaning hinges on the image to resolve references and details. Other types describe relationships where the caption carries most of the meaning or where context outside the image helps interpret it, rather than the image itself being the primary source of meaning.

Grounding between caption and image means the visual content is essential to understand the caption. In a picture-specific relationship, the image provides the concrete details that give the caption its full meaning—without seeing the image, the caption would be vague or ambiguous. The image supplies who, what, where, and how, so the caption only makes sense when paired with that visual context.

For example, a caption like “a person with a bright umbrella” becomes clear and specific when you see the person, the umbrella’s color, and the scene. The meaning hinges on the image to resolve references and details.

Other types describe relationships where the caption carries most of the meaning or where context outside the image helps interpret it, rather than the image itself being the primary source of meaning.