Content‐based image retrieval methods and professional image users

Abstract

This article reports the findings of a qualitative research study that examined professional image users' knowledge of, and interest in using, content-based image retrieval (CBIR) systems in an attempt to clarify when and where CBIR methods might be applied. The research sought to determine the differences in the perceived usefulness of CBIR technologies among image user groups from several domains and explicate the reasons given regarding the utility of CBIR systems for their professional tasks. Twenty participants (archaeologists, architects, art historians, and artists), individuals who rely on images of cultural materials in the performance of their work, took part in the study. The findings of the study reveal that interest in CBIR methods varied among the different professional user communities. Individuals who showed an interest in these systems were primarily those concerned with the formal characteristics (i.e., color, shape, composition, and texture) of the images being sought. In contrast, those participants who expressed a strong interest in images of known items, images illustrating themes, and/or items from specific locations believe concept-based searches to be the most direct route. These image users did not see a practical application for CBIR systems in their current work routines.

In a Perfect World: Exploring the Desires and Realities for Digitized Historical Image Archives

“…Interviewees were intrigued by the possibilities CBIR techniques can offer, such as automatic character recognition. This finding conflicted with the results by Beaudoin (2016) who argued that scholars in archaeology and art history were not interested in CBIR systems but wanted to rely on textual descriptions. However, it is likely that during the six years between the studies, technology has developed and users are more familiar with the supports AI can offer.…”

Billedindeksering og sociale medier

“…Viden om CBIR er nødvendig, og undersøgelserne viser, at brugere, som er interesseret i disse systemer, primaert er interesserede i søgning efter formelle karakteristika i fotografier. (Beaudoin, 2015;Mounika et al, 2016) Artiklerne i den tredje sub-kategori, Queries, fokuserer på design og udvikling af effektive fotosøgesystemer baseret på aktuelle søgninger efter digitale billeder. Spørgsmålsformulering og specielt reformulering er et vanskelligt område, og et tesaurussystem er blevet foreslået etableret til visning af synonymer og hierarkiske termer, som kan hjaelpe brugerne til at konstruere deres søgninger mere korrekt og effektivt.…”

Understanding subjects contained in Dunhuang mural images for deep semantic annotation

“…concerning images, set up an image application framework about these users, included the demand and methods while they use images, finding that people with different occupations have different motivations and approaches. Based on this, a content-based image retrieval (CBIR) system was established, and illustrated the use-context of the CBIR (Beaudoin, 2016).…”