Modeling and analyzing the topicality of art images

Abstract

This study demonstrates an improved conceptual foundation to support well‐structured analysis of image topicality. First we present a conceptual framework for analyzing image topicality, explicating the layers, the perspectives, and the topical relevance relationships involved in modeling the topicality of art images. We adapt a generic relevance typology to image analysis by extending it with definitions and relationships specific to the visual art domain and integrating it with schemes of image‐text relationships that are important for image subject indexing. We then apply the adapted typology to analyze the topical relevance relationships between 11 art images and 768 image tags assigned by art historians and librarians. The original contribution of our work is the topical structure analysis of image tags that allows the viewer to more easily grasp the content, context, and meaning of an image and quickly tune into aspects of interest; it could also guide both the indexer and the searcher to specify image tags/descriptors in a more systematic and precise manner and thus improve the match between the two parties. An additional contribution is systematically examining and integrating the variety of image‐text relationships from a relevance perspective. The paper concludes with implications for relational indexing and social tagging.

Rethinking image indexing?

“…Taking a closer look at Huang et al's () argument, it should be noticed that Pointillism is described as an artist's technique at the same time that style is part of the separated visual elements in Layer 1. This is confusing: Is Pointillism a technique?…”

“…In the following, three aspects are in concern: (a) the use of an interpretation model created by the renowned art historian Erwin Panofsky (1892–1968); (b) the danger of simplifying theoretical approaches from domains outside information studies, in this case approaches from an art historical methodology; and (c) attention to new “mutations” that might override the old ways of doing things. In particular, the article “Modeling and Analyzing the Topicality of Art Images” (Huang, Soergel, & Klavans, ), recently published in this journal, will be targeted as an indication of these aspects.…”

“…In defining the upper iconological sublayer (2.3), Panofsky is cited explicitly: the intrinsic meaning “… ‘is apprehended by ascertaining those underlying principles that reveal the basic attitude of a nation, a period, a class, a religious or philosophical persuasion – qualified by one personality and condensed into one work’ …” (Huang et al, , p. 1618). This, however, is added on immediately following: “… speaking from the Perspective 1, Art history and theory with emphasis on the artist's intent.” This perspective is, in Huang et al's () overall theoretical approach, supplemented by “Perspective 2', ‘Viewer's perception. Emphasis on viewer's response’ ”; both perspectives, constituting the so‐called “Facet 2: Perspectives on image topicality,” can be applied to each level of “Facet 1: Layers of image topicality” (Huang et al, , p. 1617).…”

“…This, however, is added on immediately following: “… speaking from the Perspective 1, Art history and theory with emphasis on the artist's intent.” This perspective is, in Huang et al's () overall theoretical approach, supplemented by “Perspective 2', ‘Viewer's perception. Emphasis on viewer's response’ ”; both perspectives, constituting the so‐called “Facet 2: Perspectives on image topicality,” can be applied to each level of “Facet 1: Layers of image topicality” (Huang et al, , p. 1617). Huang et al () also present a “Facet 3: Image topicality relationships,” which, although important for their theoretical approach, is irrelevant in the present critique.…”

“…Emphasis on viewer's response’ ”; both perspectives, constituting the so‐called “Facet 2: Perspectives on image topicality,” can be applied to each level of “Facet 1: Layers of image topicality” (Huang et al, , p. 1617). Huang et al () also present a “Facet 3: Image topicality relationships,” which, although important for their theoretical approach, is irrelevant in the present critique.…”