Image Captioning
Definition
Image captioning generates natural text descriptions for the visible content of an image.
Models combine computer vision and language generation to convert objects, scenes, or relationships into text. A generated caption is not automatically a suitable accessible alt text.
Simply explained
Like a system that describes in one sentence what it recognizes in an image.
Why it matters in practice
Relevant for media organizations, search, and as an aid for accessibility – with human review in the publication context.
Example from business
A DAM system suggests image descriptions; an editorial team reviews them and formulates context-appropriate alt texts.
Limits and pitfalls
- Equating caption and alt text; hallucinating people or attributes; ignoring the page context; automatically describing and publishing sensitive images.
Origin & history
Early deep learning systems combined CNNs and RNNs for image descriptions; Vinyals et al. introduced "Show and Tell" in 2014/2015.
Use in mid-sized companies
- Possible applications: media metadata, image search, and editorially reviewed suggestions for image descriptions.
Frequently asked questions
Is a caption automatically good alt text?
No, alt text depends on purpose and context. Can the model invent details? Yes.
Sources
- Show and Tell · Vinyals et al., Show and Tell
- Images Tutorial · W3C WAI, Images Tutorial
Sources checked on 10.10.2026.
Editorially reviewed on: