Human Interface for Multimedia Database with Visual Interaction Facilities

Summary

This paper describes visual interaction mechanisms for image database systems. The typical mechanisms for visual interactions are query by visual example (QVE) and query by subjective descriptions (QBD). The former includes a similarity retrieval function by showing a sketch, and the latter includes a sense retrieval function by learning user’s personal taste. We modeled user’s visual perception process by lour levels; a physical level, a physiological level, a visual psychological level, and visual cognition level. These models are automatically created by image analysis and statistical learning, are referred as multimedia indexes to database systems.