visually similar imgs

Rebecca Lieberman

Demented Panda and Koki wandered through the small plot of land. Except it was no longer only a small plot of land, but also an enormous food court. Except it wasn’t just a food court, but also an outdoor rehearsal space lent to artists by a small nonprofit arts organization. Except it wasn’t a rehearsal space, but a soundstage for gigantic live entertainments. Except it wasn’t a soundstage, but a fake Baghdadi neighborhood… Not a fake neighborhood but an intersection in the Financial District on the night of March 22, 2003. Not an intersection but an interrogation room. Not an interrogation room but a holding cell funded by the Department of Homeland Security… Not a warehouse-turned-holding cell but a warehouse-turned-club where the after-party takes place. Not an after-party but an academic conference on politics and aesthetics. Not a conference but a boardroom meeting on tax-deductible philanthropic donations to nonprofit arts organizations. Not a boardroom but a bunker, dug into the wet and muddy ground. [1]

visually similar imgs is an ongoing multimedia investigation into collecting, archiving and reassembling digital material from Google’s Search by Image. Gesturing at the seams of familiar technology, the project uses images mined from search results in which Google fails to match an image based on “visual DNA.” Through image sequences that take multiple forms — a series of artist publications, looped video projections and a net-based image work — visually similar imgs explores the poetics of search by subjecting the algorithmic logic of reverse image searching to affective and artistic response.

Google Search by Image allows users to reverse-search images using visual data instead of written queries. Though Google keeps the specifics of its algorithm somewhat obscured, it appears to work by analyzing the image itself to assemble a “visual DNA” of characteristics for use in matching it with other images. Though the algorithm is sometimes successful in its design (allowing users to discover the virtual provenance of an image), its frequent moments of failure are both peculiar and interesting. As the seams begin to show themselves (sometimes in ways that are hilarious, ridiculous, or even grotesque), they afford a glimpse of the poetic.

When Search by Image can’t match an image exactly, it generates a proliferation of new images that are “visually similar.” Images become material: shapes and forms, maps of colors and brightness, pixel topographies, histograms of the everyday. They take on the ability shape shift into other things—a scene from a sci-fi film apocalypse parades as a fiery sunset and a fluorescent hotel room.

visually similar imgs taps into the poetics of search and engages with poetics in a few important ways. It seeks to make the Internet’s detritus expressive, to transform its bottom feeders into a kind of visual poetry. It liberates them from conventional structures of meaning and reconstitutes them as semantic building blocks. It subjects the algorithmic logic of search to affective response. It looks for the poetic among the algorithmic, the subjective among the automated. It sees ripe and radical potential in the moments where technology fails, in seeing as a computer might see. It makes a book into a dreamscape, and a browser into a hall of mirrors where objects endlessly mutate and multiply. And like the protean space of Demented Panda and Koki, in which theater is a warehouse is a bunker, these images inhabit a universe in which things become other things, depending on how you look at them.

So where were Demented Panda and Koki?

Not a boardroom but a bunker, dug into the wet and muddy ground. Not a bunker but a book, each line redacted except for the numbers. Not a book but a bonfire made from its burning pages. Not a bonfire but a set of bright stage lights, illuminating the small plot of land so that the audience could better see the action.[2]

Lieberman_Media N Layout2_101014-1

Lieberman_Media N Layout2_101014-2

Lieberman_Media N Layout2_101014-3

Lieberman_Media N Layout2_101014-4

Lieberman_Media N Layout2_101014-5



  1. David Buuck and Julianna Spahr, An Army of Lovers (San Francisco, CA: City Lights, 2013), 33.
  2. Ibid.



Rebecca Lieberman is a Brooklyn-based artist, designer and creative technologist. She works across a range of disciplines and materials, such as videos, sculptures, artist publications, and net-based work. Rebecca earned her undergraduate degree in 2010 from Harvard’s Visual and Environmental Studies department, where she learned from Fluxus artists, and completed a thesis in sculpture and video installation. She has shown her work at Anthony Greaney Gallery in Boston, Human Resources in Los Angeles, the Harvard Museum of Natural History, The Carpenter Center for the Visual Arts, and Galerie der HfK in Bremen. She is currently a graduate student in NYU’s Interactive Telecommunications Program.