A Multi-faceted Browsing Interface for Digital Photo Collections

摘要

Managing photos by using visual features (e.g., color and texture) is known to be a powerful, yet imprecise, retrieval paradigm because of the semantic gap problem. The same is true if search relies only on keywords (or tags), derived from either the image context or user-provided annotations. In this paper we present a new multi-faceted image search and browsing system, named Scenique, that allows the user to manage her photo collections by using both visual features and tags, possibly organized into multiple dimensions (or facets). Each facet can be seen as a coordinate of a multidimensional space describing the image content (for example, the visual appearance, the content type, the geographic location, and so on). We present the basic principles of Scenique and provide evidence of the effectiveness of its visual tools. Feedback supplied by a set of real users indicates that the proposed interface is intuitive, easy-to-use, and that satisfies users' expectations in managing photo collections and quickly locating images of interest.