Google had introduced an algorithm called PageRank nearly 10 years ago which changed the way users search for web pages. Now, Google has innovated mechanism for its image search. Google says that it now has a technology that will have the same impact on online image search. At the International World Wide Web Conference held in Beijing recently, two researchers from the Google presented a paper called ‘PageRank for Product Image Search’, announcing a fledging algorithm that renovates the traditional text-based methods employed by the company's current image search technologies.

 

Shumeet Baluja and Yushi Jing, who presented the Google paper, said that their experiment results had resulted in a vital improvement in terms of user satisfaction and relevancy when compared with the present Google Image Search output. The truth, infact, is that that the current Google image search service is really lousy. When sorting out images, the present search engine hardly gives any time on scrutinizing the pictures. Instead, the results shown as a result of image search, are on the basis of the text surrounding the image.

 

Google's new PageRank for Product Image Search that is also known as ‘VisualRank’, by contrast, seeks to actually comprehend what's pictured. For instance, researchers Shumeet Baluja and Yushi Jing showed an image search on the word ‘McDonald's’ using the new technology. In this case, VisualRank identified the famous golden arches as theme. An image dominated by the golden arches would then be ranked higher than another picture where the arches are tucked into the background.