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 percent less irrelevant images.
So, has Google finally figured out how to index images without relying on the surrounding text or image file name? Reading the research paper would suggest it has.
In terms of overall performance on queries, the proposed approach contains less irrelevant images than Google for 762 queries. In only 70 queries did Google’s standard image search produce better results. In the remaining 202 queries, both approaches tied (in the majority of these, there were no
And Google gives an example of just how accurate the new image search is…
However, the key question is can such a model scale? Can it be applied to the billions of images floating around the web? Munjal Shah the chief executive of Riya–a search engine that matches colors and shapes–doesn’t think it will scale.
“I think what they’re trying to accomplish is largely impossible,” he said. “Our belief is, there is not large-scale solutions.”
“Impossible” you say? That sounds like a gauntlet that Google will happily take up!