Google just announced yet another new product in Google Labs: Google Image Swirl.

Image Swirl organises image search results into groups and sub-groups, based on their visual and semantic similarity and presents them in an intuitive exploratory interface. It is an altogether new way of searching for images which could help resolve an ambiguous query visually (apple, jaguar, beetle).

Google also hopes that the Wonder-Wheel-like user interface could also provide inspiration to explore a concept from different visual perspectives (Eiffel Tower, beach, impressionism).

Rather than just showing one image, the Image Swirl feature displays a whole range of photos. Clicking on this stack opens up a Wonder Wheel with related images clustered around the original photo (not unlike the Wonder Wheel search feature recently embedded into YouTube).

For now, Image Swirl only works for about 200,000 queries, though Google plans to include more in the near future.

To create these clusters of related images, Google analyzes the characteristics of the images themselves, but the algorithm also looks at meta tags and other clues in the description of these images. Google uses the same algorithms to find and organize images of landmarks in its index.

Google Image Swirl Product Manager Aparna Chennapragada said that this is new service is part of Google drive to “go beyond just relying on text.”

Bing introduced its visual search feature a few weeks ago and is still ahead of Google in bringing these features to its core search product.

Bing's visual search

So far, the comments have been fairly positive:

Comments from the Image swirl Google Labs page

[thanks to ReadWriteWeb for the tip-off]