Back in 2001, to give people a new, quicker way to find images, we launched Image Search. When you do a search for [eiffel tower] you’ll find an array of images of the tower in the daytime, in black and white, at sunset and more. With Similar Images, which recently graduated from Google Labs, you can click “Find similar images” to narrow your search to, say, pictures of the Eiffel Tower lit up at night. Today, we’ve launched an experimental feature in Labs called Google Image Swirl, which builds on new computer vision research to cluster similar images into representative groups in a fun, exploratory interface.
A fun, exploratory interface? Sounds like some kind of alien probe in a not so pleasant area but hey it’s engineers talking here, right?
Here’s the initial page you encounter when doing an image search in Swirl.
Then you can further break down your images to see them like this
Image Swirl expands on technologies developed for Similar Images and Picasa Face Recognition to discern how images should be grouped together and build hierarchies out of these groups. 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.
So it’s neat but it’s not anything that will change the face of search. Hasn’t Google done enough of that already?