VisualRank is "an algorithm for blending image-recognition software methods with techniques for weighting and ranking images that look most similar," writes the New York Times. It tackles some fundamental issues that make image searches problematic. Right now, image searches turn up results based on the keywords that are tagged to the photos, not the objects in the photos themselves. So you may be looking for a picture of that sexy new HDTV from Philips, but instead you get a picture of Christian Slater because he's tagged as providing the voice over for the TV's commercial. Not a very useful result.
In response, Google has begun indexing images on the Web. Not all of them, mind you. That task is deemed nearly impossible. The New York Times explains:
With 83% less irrelevant images, it's obviously far from perfect. This should still reduce the time it takes to find the exact image you're seeking. And as we all know, speediness is one of Google's fortes.
Google does not disclose how many images it has cataloged, but it asserts that its Google Image Search is the "most comprehensive image search on the Web." The company said that in its research it had concentrated on the 2000 most popular product queries on Google's product search, words such as iPod, Xbox and Zune. It then sorted the top 10 images both from its ranking system and the standard Google Image Search results. With a team of 150 Google employees, it created a scoring system for image "relevance." The researchers said the retrieval returned 83% less irrelevant images.