Posted April 28, 2008 10:30 am by with 7 comments

Tweet about this on TwitterShare on LinkedInShare on Google+Share on FacebookBuffer this page

Google has a lofty ambition when it comes to image search. The world’s largest search engine hopes to do for image search what it did for regular text search–make it a whole lot more reliable.

How it plans to do this is outlined in a new research paper written by two Google scientists. According to the NYT:

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
irrelevant images).

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!

  • Pingback: Google Discovers Holy Grail of Image Search, But Will it Scale? - Marketing Pilgrim | Zune()

  • Yeah, just leave it to Google to do the impossible.

    Eric Martindale’s last blog post..Today’s My Birthday. Want A Link?

  • I haven’t read the research paper yet, but it’s downloaded and waiting for me. I have a hard time seeing how Google can pull this off, but it will be great if they can.

  • Personally,
    I dread the thought of improved image search.
    The only accomplishment that is achieved by such is more illegal bandwidth theft. The fact is, anyone needing an image for legitimate use will have sources other then Google’s image search to accomplish their task at hand.
    Heidi-Ann Kennedy
    Scientific Frontline

  • @Heidi-Ann – that is a legitimate concern. Whenever clients ask me about optimizing their images, I always warn that it encourages hot-linking.

  • Yes,I believe it will scale and if anyone can do it, Google is going to make the nay-sayers have to drink a lot of Pepto ’cause they are going to be eatng their words. Yes, I believe Google can do it and do it Well!

  • This does seem ambitious. And are most images really worth being indexed?