"Dreamscapes" by Daniel Ambrosi
By Anna Marks, contributor
Daniel Ambrosi is a California-based artist whose passion and work takes him all around the U.S., including Colorado, Half Moon Bay, Utah, and Lake Tahoe. Since graduating from Cornell University in 3D graphics and architecture, Ambrosi has been exploring new ways of visual presentation using various digital technologies. In this current body of work, he uses computational photography to shoot panoramic landscapes, resulting in eye catching and vibrant imagery. Specifically, Ambrosi modifies a version of "DeepDream," Google’s vision program, which was developed to visualize the inner workings of artificial intelligence models. By utilizing and manipulating this technology, Ambrosi constructs psychedelic and dazzling artworks which display the fine details of landscapes that are not always captured using standard photographic techniques. Ambrosi’s work illustrates that experimenting with technology to manipulate photography can give natural landscapes - as naturally beautiful as they are - an intriguing new aesthetic and narrative.
Daniel Ambrosi (center) and one of his two engineering collaborators, Joseph Smarr (left), interactively explore their “Dreamscapes” on a giant display system at the Qualcomm Institute in UC San Diego on March 31, 2016. This was the first time the artist and engineer were able to see their multi-hundred megapixel creations at lifelike scale.
The Azalea Walk Dreamscape is composed of 63 individual photos (7 wide x 3 high x 3 “deep”) that are seamlessly stitched and blended together using Ambrosi’s “XYZ” computational photography technique. This 270+ megapixel image was captured and assembled in 2013 and then transformed using Ambrosi’s modified “DeepDream” artificial intelligence (AI) software in 2016. The zoomed-in detail on the right shows the surprising details created by the AI, which are only visible from close range.
Dreamscapes made their public debut on April 1, 2016, at the annual GPU Technology Conference in San Jose, CA, where conference host, NVIDIA Corporation, featured them as a series of 8’ high x 16’ wide backlit tension fabric structures. Since these images depend on NVIDIA graphics processing units (GPUs) to run the artificial intelligence program Ambrosi is using, NVIDIA purchased these pieces to showcase the power of their technology.