Improving user interaction with image collections

To date, research in image retrieval has focused on the underlying retrieval algorithms, while little research has been focused on the many aspects pertaining to user expectations, user actions and user experiences. It is here postulated that ultimately, it is the user who decides the quality of an image retrieval system based on a match between his/her expectations and system performance. This paper presents a study of user expectations and experiences from query formulation in image retrieval. The study used an image retrieval prototype that supports text-based and content-based image retrieval, as well as a combination of the two. Results indicate that users are able to make use of a range of image retrieval tools, that content-based query formulation is not well understood, and that the different approaches to image retrieval may be useful in different situations depending on if the users know what they are looking for.