Bildfähig!

1

Among the crucial challenge of any ambitious expert in contemporary visual culture are decisions of how he will allow his thought on pictures be affected by the challenges of digital image technology, if he uses it for research. Content based image retrieval systems (see here, here and here), which have improved strongly over the years, may be at the core of the answers to this question. In short, these systems allow to search pictures after visual patterns respectively pictures themselves and not simply by text. It would be particularly surprising, if anybody studying visual phenomena by contemporary image makers, would not use a system, that allows to look at pictures ‚through’ other pictures.

At a conference a couple of years back, Claus Pias has assumed that image experts in the future will be primarily concerned with the „infinite possible and impossible configurations of data“ resp. images. Thus, the challenge of any visual analysis will be to use the possibility to search by visual patterns in a most intelligent and illuminating way, i. e. any researcher of contemporary visual phenomena will try to improve his abilities in getting to interesting relationships between pictures.

Among the many changes that this may lead us to, I think, the gain of importance of one certain kind of analysis will be the most interesting. In an old but still very useful article on the problem of theories that are based on the dichotomy of high and low culture texts, literary critic Helmut Kreuzer argues in favor of a vertical rather than horizontal, i. e. dichotomic mode of analysis. Such a vertical analysis avoids to rank texts in high and low, but rather focuses on the analyses of differences between them. This is certainly applicable to pictures as well and in fact, it is less and less compelling, if not undesireable for anybody concerned with contemporary visual phenomena to base one’s thinking on the dichotomy criticized with precision by Kreuzer.

I believe that content-based image retrieval systems will respond very positively to any attempt to pursue such a vertical mode of analysis. The reasons are obvious: As pictures become searchable by pictures, we will be able to look more and more into visual patterns and characteristics common to a certain group of pictures. We will be able to understand much better the relationship between pictures and as a result how identical or similar visual patterns and modes of visual representation can be used in different contexts answering to different expectations.
A part from these future possible endeavors, it is hard to guess how image production will change, if search engines will offer such a system to all users. And it is equally hard to guess, how many users will be interested . If it gets also used by non-professionals, image experts may face another challenge. They would have to not only deliver intelligent remarks on pictures, but also ones that are applicable to the new search opportunities and produce new and interesting groups of search results. Not only will they have to be competent in explaining and analyzing individual details of pictures, but also thinking more about how their remarks will be able lead their readers to other interesting pictures.

Bildquelle: 1 screenshot, 26.11.2008, detail, CIRES website.

References:

More on content-based image retrieval, here.

Helmut Kreuzer. “Trivialliteratur als Forschungsproblem. Zur Kritik des deutschen Trivialromans seit der Aufklärung,” In: Deutsche Vierteljahresschrift für Literaturwissenschaft und Geistesgeschichte Bd. 41 (1967): 173-191.

Claus Pias. “Das digitale Bild gibt es nicht - Über das (Nicht-)Wissen der Bilder und die informatische Illusion” In: zeitenblicke 2 (2003), Nr. 1. hier.

More English posts: here.