Perhaps there are images on a client's site that they'd rather not have indexed, either because they draw in the wrong kinds of traffic, increase bandwidth costs or strain the servers. Philipp Lenssen
outlines the steps you can take to prevent the engine's crawlers from accessing those images, with the first strategy to include the "noimageindex" tag in the header portion of a page's HTML.
But the "noimageindex" meta data may not be enough if other Web sites can still link to those images (as the crawler will be able to access them through the link). So Lenssen suggests
creating a separate server folder for all of the images and then making sure to disallow the crawlers access to that folder via the robots.txt file.
Read the whole story at Google Blogoscoped »