Journal of the American Society for Information Science · 1999 · 48 citations · 7 references
ComputingComputer ScienceInformation RetrievalNatural Language ProcessingVisual RepresentationConversational InterfacesConversational User InterfacesDialogue SystemsImage RetrievalMultimodal InteractionTextual RepresentationCaptioningCaptioning AccessImage CaptioningImage-text InteractionImage-text RetrievalRetrieval-based Response GenerationText-to-image GenerationUser-guided Generation
Words are problematic for describing images, but they are a convenient and traditional way of describing requests. Users can given voice to their reactions to images-how well they suit needs. User-generated reactions might provide word-based descriptors helpful to subsequent users and requiring minimal system resources to produce. Shifting focus from description of documents to description of reactions is accomplished by gathering verbal captions and responses to images. User generation of captions and verbal responses within a collection of 300 diverse images is demonstrated and analyzed. Functional adjectival descriptors appear in 20% of the responses and functional narrative (conversational) descriptors appear in 80% of the responses. Issues of larger scale analysis, implementation, and possible shifts in understanding of representation for retrieval are discussed.
7
Analysis of user need in image archives
Linda H. Armitage, Peter Enser · Journal of Information Science · 1997 · 248 citations
Raya Fidel · Journal of the American Society for Information Science · 1994 · 85 citations