San Francisco: Researchers at Google Inc. say they have a software technology intended to do for digital images on the Web what the company’s original PageRank software did for searches of Web pages.
On Thursday at the International World Wide Web Conference in Beijing, two Google scientists presented a paper describing what the researchers call VisualRank, an algorithm for blending image-recognition software methods with techniques for weighting and ranking images that look most similar.
Although image search has become popular, results are usually generated today by using cues from the text associated with each image.
Despite decades of effort, image analysis remains a largely unsolved problem in computer science, the researchers said. For example, while progress has been made in automatic face detection in images, finding other objects such as mountains has lagged.
“We wanted to incorporate all of the stuff that is happening in computer vision and put it in a Web framework,” said Shumeet Baluja, a senior staff researcher at Google, who made the presentation with Yushi Jing, another Google researcher.
The company’s expertise in creating vast graphs that weigh “nodes”, or Web pages, based on their “authority” can be applied to images that are the most representative of a particular query, he said.
The company said that in its research it had concentrated on the 2,000 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.
Google is not the first into the visual product search category. Riya, a Silicon Valley start-up, introduced Like.com in 2006. The service, which refers users to shopping sites, makes it possible for a Web shopper to select a particular visual attribute, such as a certain style of brown shoes or a buckle, and then be presented with similar products available from other Web merchants.
Rather than relying on a text query, the service focuses on the ability to match shapes or objects that might be hard to describe in writing, said Munjal Shah, the chief executive of Riya.
“I think what they’re trying to do is largely impossible,” he said. “Our belief is, there is no large-scale solutions.”
Shah said there had been a number of technology demonstrations by Google Labs researchers that used machine-learning techniques to recognize a person’s gender. However, the company has been slow to deploy its research.
©2008/THE NEW YORK TIMES