Image attributes: A study of scientific diagrams

Abstract

AbstractIn recent years, advancements in imaging technology have fostered a dramatic increase in user access to digital images. In accordance with this trend, greater emphasis has been placed on efforts to develop adequate indexing and retrieval methods for image databases. However, lacking a detailed basis upon which to derive such methods, relevant efforts are often developed based upon an inconsistent set of general theories and assumptions regarding user needs and search strategies. Citing the realistic need for an empirical basis to address these issues, recent efforts have focused upon human subject testing as a means to ascertain the general attributes humans use when describing certain images. The following paper details the preliminary results of a relevant effort to explore the attributes naïve subjects use to describe scientific diagrams. Results of this investigation are intended to expand the current subset of images explored through this methodological approach, as well as explore the importance of certain image attributes across different types of visual representations.