Search, Browse & Curate: 27 August 2014

A recent study by researchers at Rutgers University has suggested an algorithmic approach towards revealing influences among paintings through the use of semantic based classification of images. Their study revealed known as well as possible influences among paintings— some of which have not been previously made by art historians.

The study stands as a foil to other approaches that are based on visual properties such as brightness or hue. These include Lev Manovich’s Style Space or the Cooper Hewitt’s color browser.

All of these experimental algorithmic approaches are in pursuit of finding new connections and new interpretations of the vast collections of images now digitized. They suggest a possible blurring of browse, search, and curation related activities that could be useful in reimagining a more useful and pleasurable way of navigating a digital library.

In fact, the American Folk Art Museum’s 2011 exhibition of quilts at the Park Avenue Armory, Infinite Variety, posits this possibility in an almost algorithmic display of visual artifacts that encodes information spatially. The experience is at once visually compelling and highly informative.

Such an approach, if applied to a digital collection, could enhance discoverability and user retention. More importantly, browse/search does not become a passive means to an end but an active experience that facilitates new interpretations of the visual world.