Finding cultural heritage images through a Dual-Perspective Navigation Framework

Abstract

With the increasing volume of digital images, improving techniques for image findability is receiving heightened attention. The cultural heritage sector, with its vast resource of images, has realized the value of social tags and started using tags in parallel with controlled vocabularies to increase the odds of users finding images of interest. The research presented in this paper develops the Dual-Perspective Navigation Framework (DPNF), which integrates controlled vocabularies and social tags to represent the aboutness of an item more comprehensively, in order that the information scent can be maximized to facilitate resource findability.DPNF utilizes the mechanisms of faceted browsing and tag-based navigation to offer a seamless interaction between experts' subject headings and public tags during image search. In a controlled user study, participants effectively completed more exploratory tasks with the DPNF interface than with the tag-only interface. DPNF is more efficient than both single descriptor interfaces (subject heading-only and tag-only interfaces). Participants spent significantly less time, fewer interface interactions, and less back tracking to complete an exploratory task without an extra workload. In addition, participants were more satisfied with the DPNF interface than with the others. The findings of this study can assist interface designers struggling with what information is most helpful to users and facilitate searching tasks. It also maximizes end users' chances of finding target images by engaging image information from two sources: the professionals' description of items in a collection and the crowd's assignment of social tags.

Using eye movement sequences and hidden Markov model to explore strategy blending in metadata-supported image search

“…Stvilia et al's [39] ndings along with a later work [40] suggest that folksonomy increases access points and supplements expert-created subject headings. Lin et al [27] also show that integrated interfaces are more e cient than both single descriptor interfaces (subject-heading-only and tag-only interfaces). Since metadata quality is subjective and context-dependent [41], the ways a user perceives it can be context-dependent and can change during the search process.…”

“…Clicking on any image lead users to a Detail (D) page containing information about the chosen image. The effectiveness and usefulness of the proposed system has been carefully studied in previous work [27].…”

“…The images collected were diverse in style, including music, sports, geography, etc. The experimental tasks were designed based on a previous work [27]. According to the results of this previous work, the present study focuses only on exploratory search, which requires continuous and iterative searching, analysis, and result evaluations.…”

“…Different from traditional researches that use mathematical models to analyze user behavior, this study attempts to investigate user search pattern through experimental methodology. Three image search interfaces are developed: a tag-based interface, a subject heading-based interface, and a dual-perspective interface [27]. This study uses an eye tracker to obtain eye movement data to lay the groundwork for examining image search with different interfaces.…”

Coding peekaboom: a gaming mechanism for harvesting programming concepts

“…Social tags not only represent the content but also bridge the gap between humans and machines by including the user aspect in the indexing task (Lee et al, 2012). They have been widely used to improve web search (Lin et al, 2016(Lin et al, , 2019, especially in terms of indexing. This is also the motive to design the proposed system, Coding Peekaboom, with tags to effectively index the code materials manually collected from StackOverflow.…”