Content-based image similarity search — determining the images in a database visually similar to a search image on the basis of pixel data alone — is a hot topic. Applications are numerous. Just recently, amazon.com has introduced a visual product similarity search for some products, and google has introduced a search-by-image feature.
In addition to hands-on applications, there is a lot of research on the topic. The abilities of university maintained image search engines grow constantly. Unfortunately, implementing current techniques requires a bunch of specialist knowledge and some mathematical sophistication. Fortunately, however, there are effective simple techniques for which there exist open source libraries.