Google takes a visual search
Theme image search already raised in these pages, when the company considered Photodate , Developers who have , And Idee Piximilar . Now getting into the heavyweights.
Google will face Yushi Jing and Shumeet Baluja conference WWW2008, held April 21-25, 2008 in Beijing has demonstrated its development (Link to PDF, 3,4 Mb), capable of analyzing images not only in name, as it was before, but the quality, relevance and most importantly - content.
Now image search on Google images work on the basis of the text, as looking at file names. And as the name can be assigned arbitrarily, depressing picture of the issue and cheers. Try a query . You will find on the first page pictures fat giraffes, fat women and fat children, which, of course, has no direct relation to the request.
Our approach relies on the analysis of the distribution of visual similarities among images.
The premise is simple: the author of a Web page is likely to choose the images that, with his or her opinion, are relevant to your topic. Instead of assuming that every user who has a web page relevant queries will link to the image on the page of another user, our approach relies on the analysis of the combined preferences of many of the creators of the content network.
The new image search on Google will work on the basis of the so-called "visual themes", which can be explained as a semantic center of the image.
For example, to illustrate many of the images contain a familiar character «M». In some emblem - the main center of the image, while in others it occupies only a small part. Nevertheless, its recurrence in a large fraction of images - an important signal that can be used to deduce the overall "visual theme" throughout the set. The discovery of multiple visual themes and their relative strength in a large set of images - image based ranking system proposed in this study.
Like the Page Rank (PR), introduces the concept of Image Rank, IR. Formula Image Rank is defined as:
where S - the normalized matrix of similarity (normalized, symmetrical adjacency ma-
trix) images, such as S u, v denotes the visual similarity between images u and v.
Google has already demonstrated operation of its graphical search engine on the example of 2000's most popular searches in Google images.
The company is close to the submission of the first beta version of a graphical search engine.
Link: (English)
Blogobzor: And you know what a cloud of blogs ?
May 4, 2008 at 22:18
In fact, it is another hype, which rises as soon as Google sneezes. If you read this publication, as well as its own company statement, it becomes clear that this is only research, and they are still very far from realization. By the way, in contrast to systems that were previously covered in this blog and who have a valid version of the engines. Most likely, eventually Google or Microsoft will buy one of these companies (or all:)), and that's when something will work.
May 4, 2008 at 23:45
Alexis, you're not from Picollator'a?
In my piercing gaze:) engines illuminated before the company's blog at the moment are far from the ideal of human search.
Ie they look all right (with t.zr. robot): the color match, and shape, but is all wrong.
Just from Google, you can expect something more serious.
All the prerequisites are. In addition, the formula of ranking all prechislennyh systems were not observed, as far as I know, but this is for Google - a priority.
Well, wait and see.