Trevor Paglen

The Black Canyon Deep Semantic Image Segments

The Black Canyon Deep Semantic Image Segments by Trevor Paglen merges traditional American landscape photography (sometimes referred as ‘frontier photography’ for sites located in the American West) with artificial intelligence and other technological advances such as computer vision. In order to take this photograph, Paglen traveled to the Black Canyon, south of the Hoover Dam. Only accessible by water, Paglen piloted a boat up the Colorado river into the canyon. Paglen was drawn to the canyon because of the significant role it played in 19th century ‘frontier’ photography. Today, this area is much less photographed than similarly popular locations because it is only accessible by boat. The artist admits that the trip to photograph this historic site wasn’t altogether enjoyable, as he went in the dead of summer while the  temperature was 125 degrees fahrenheit. During his stay, he camped along the riverside and barely slept due to the extreme heat. This experience gave Paglen a greater appreciation for the grueling conditions that frontier photographers worked under during the 19th century. 

The intensely saturated color in the image is due to Deep Saliency, which Paglen explains is produced by “using computer vision to analyze photographs using Artificial Intelligence. In ‘classical’ computer vision, one develops algorithms that look for lines, circles, ‘interesting features,’ and complex shapes and tries to infer something about the images from these ensembles […]  one ‘trains’ the neural network on thousands or millions of images, and the network develops its own ‘tools’ to analyze those images. Deep Saliency is a technique to differentiate between sections, areas, or different types of objects in an image as interpreted by a neural network using criteria it has created for itself.” (Altman Siegel, Trevor Paglen, 29).