Google image search now lets you 'swirl'

Google is testing a new experiment in its image search function, allowing images to be layered onto various clusters. Google Image Swirl blends images into "representative groups in a fun, exploratory interface." When you look for pics on Google Image, you often get a boring palette of images that may or may not be related to your search term. Finding relevant photos to "Washington" may be difficult, for instance, because Google will display the White House, the Washington Monument, the Washington Redskins football team, and more. Now Google wants to "discern how images should be grouped together and build hierarchies out of these groups," according to its recent blog post. Today, in the Labs section, Google is testing Google Image Swirl, an interactive alternative to the usual search function. Essentially, Swirl creates clusters and sub-clusters of the search term. Click on a thumbnail of Johnny Depp and it'll expand to show him in different scenes or in different positions, such as headshots or movie stills. The clusters are grouped together based on metatags. As Google explains in a somewhat verbose manner: "Each thumbnail on the initial results page represents an algorithmically-determined representative group of images with similar appearance and meaning. These aren't just the most relevant images — they are the most relevant groups of images."

From Google Image Swirl

Searching for Toronto Blue Jays on Swirl

image:59629:2::0

Swirl provides an intuitive bird's-eye view of what photos you want to select. Dividing images into sub-groups is especially useful for finding sub-categories of certain terms. For instance, if you look for "toronto blue jays", Swirl will divide the images into clusters, some pooling together images of players or the Rogers Centre stadium or baseball caps. Image Swirl currently works for more than 200,000 queries and plans to include more queries in the future.