arXiv.itsnick.co

an independent field guide to 23 million papers from arXiv & OpenAlex — not affiliated with arXiv.org

Computer Science · Computer Vision and Pattern Recognition

n Adaptive Approach for Image Organisation and Retrieval

Jana Urban

See on the map

Abstract

Image retrieval is an intrinsically hard problem. Manual labelling is impractical for most purposes and automatically extracted content-based features do not describe what humans recognise and associate with an image, referred to as the semantic gap. The semantic gap complicates the query formulation process for the searcher. Moreover, image meaning is subjective and contextdependent. Finally, information needs are often vague and dynamic, since image searches are usually coupled with creative tasks. These problems render current image retrieval systems difficult to use. Unlike most previous work in the field, which has studied the retrieval system as a self-contained problem, the approach described in this dissertation takes a holistic view, in which information access is considered as part of a larger work process. By taking into account the design of both the interface and retrieval algorithms, all of these intrinsic issues of image retrieval are addressed together. The starting point in creating a more user-friendly system was to redesign the interaction process between user and system. Based on analysing user studies—both from previous work in the literature and those described within this document—the organisation of information has been found to help structure the thought process of the searcher. Therefore, the proposed system,

More like this