Msg#: 3644387 posted 2:15 am on May 8, 2008 (gmt 0)
Abstract A system, method, and computer-readable media are disclosed for providing images in a ranked order. The system can include an aggregation component for aggregating a plurality of images with corresponding text. Additionally, the system can include a name detector a name detector for detecting names within a search query. Moreover, the system can include a ranking component for ranking the aggregated images based on whether the name detector detects a name.
For those of you involved with Image Optimization, the above patent recently awarded to Microsoft sheds a little bit of light on how to get your images to rank. After reading that multiple times, I'm beginning to wonder if allowing image hotlinking is not such a bad idea. ;)
Bill Slawski at SEO by the SEA has a pretty good writeup about it and other relevant image optimization information.
Msg#: 3644387 posted 6:57 am on May 8, 2008 (gmt 0)
I'm beginning to wonder the same thing about the hotlinking of images. :)
It was good to see aspects of image optimization discussed that go beyond the analysis of text on the same page as the pictures, or anchor text in links that might be pointing to images.
I've been rethinking some other aspects of how images are used on a site and how a search engine might determing a ranking for those images, such as what the relationships between thumbnails and the larger images they point to might tell a search engine.
Msg#: 3644387 posted 1:00 pm on May 8, 2008 (gmt 0)
There was something else in that patent that stood out to me. I almost posted a topic on it a couple of years ago in regards to Google and it has to do with image size. One of the things I've consistently noticed when doing image search are that the sizes are typically larger in the first set of results. It is not like that 100% of the time but enough to indicate to me that size does matter. ;)
Msg#: 3644387 posted 2:27 pm on May 8, 2008 (gmt 0)
There's some discussion of file size, and other aspects of pictures that don't appear on the surface to have much to do with a determination of relevance.
Those may be more of a "quality" issue, with a high quality image being considered a better option to show to a searcher than a lower quality image.
It's also possible that there's some threshold between too large and too small, when it comes to presenting pictures in image search. I haven't paid much attention to absolute file sizes when doing image searches, but I don't recall too many pictures being served to me that were extremely large, or extremely small.