Exploring users' video relevance criteria—A pilot study

Abstract

Relevance is one of the most fundamental concepts in information science. However, among the large body of relevance literature, very little of the research studied users' relevance judgments when concerning video. The pilot study described in this paper conducted exploratory work in this direction. The time-line interviewing method proposed by Dervin (1992) was applied in the study due to its exploratory and descriptive nature. Four participants were interviewed in early 2003, including a professor in communication studies, an art professor, a news video librarian and a video editor. Using inductive content analysis methods, three categories of relevance judgment criteria were summarized: textual (e.g., topicality, recency, and nationality), visual (e.g., cinematography, objectslevents, and style) and implicit (e.g., interest, familiarity, and appropriateness). Topicality was still considered the most important criteria for video relevance judgments, however, users also liked to see visual surrogates, especially those surrogates that contained motion. A more formal study is planned, and we expect that the results will not only enrich the relevance literature but also have implications for video indexing and retrieval research. IntroductionThe rapid development of digital video retrieval techniques such as, audiohisual indexing and abstraction and query by example formulation, have resulted in highly accessible digital video libraries. However, the challenges are more than just technical. There is another crucial part missing here, the users. To create a system that effectively supports users, it is essential to examine the users' needs, preferences, and work contexts (Payette, 1998). Although library and information professionals have already fully recognized the importance of studying the nature of users' information seeking behaviors and produced much related literature, few of these works are explicitly related to video searching. There is also a large body of literature discussing how people make relevance judgments when searching for textual information, but very few of these works focus on searches conducted to retrieve audiohisual materials. This situation could be explained by the lack of digital video libraries as a test bed to study people's information seeking behaviors and relevance judgment processes.Relevance is one of the fundamental and central concepts in information science. Relevance-based measures --recall and precision, are the two most commonly used criteria to evaluate the effectiveness of information retrieval systems. In addition, relevance is also a necessary part of understanding human information behaviors (Schamber, 1994, p.36). Therefore, research on users' video relevance judgment can not only enrich the literature of video information seeking literature, but also give implications to the design of effective video retrieval systems and browsing interfaces to serve users' needs. As Shatford (1994) notes, relevance criteria suggested from empirical user studies can be regarded as po...

Influences of users' familiarity with visual search topics on interactive video digital libraries

“…Although studies investigating users and information needs are found most notably throughout textual and image IR research, analyzing influences of needs and search topics for the purposes of video retrieval is far less common. This study asserts that examining users' familiarity with visually oriented video search topics, for the purposes of enhancing understanding about retrieval and system development, warrants independent evaluation, as the complexity of video can produce information needs that comprise characteristics not typical in textual or even image queries (Albertson, 2010;Yang, 2005).…”

Toward a conceptual framework of key‐frame extraction and storyboard display for video summarization

“…Participants' answers were analyzed on the basis of their falling into one of three categories (textual, visual, and implicit) identified by Yang and Marchionini (2004); the textual category includes overlay text and symbols; the visual category includes objects, events/actions, and associations between key frames; and the implicit category includes interest or familiarity.…”

How do health care professionals select medical images they need?

“…In order to access a comprehensive list of criteria, we used the lists of criteria collected by Schamber (1994) 2 ; Mizzaro (1997) 3 and Maglaughlin and Sonnenwald (2002) 4 . We also used the relevance criteria for images identified by Hung et al (2005); Markkula and Sormunen (2000) and Choi and Rasmussen (2002), and video relevance criteria documented by Yang (2005) 5 . Unfortunately not all authors defined the relevance criteria.…”