For several years, the only method for finding images on the Web has involved browsing numerous webpages, specifically using well-known search engines and directories like Google and Yahoo! Search results are generated by an automated web crawler (spider, robot, or bot), which examines the contents of its index and searches for file names and image file extensions. The HTML source code examples mentioned in the previous discussion of file formats (hyperlink references, image source tags, and alternate tags) are analyzed by web crawlers and matched to users' search query terms.
General Search Engines are Very Good
Two general search engines which offer simple to sophisticated image searching options include:
AltaVista Image Search
Allows users to search the Web for photos, graphics, buttons/banners, color and/or black-and-white images from the Web or from the websites of various affiliated companies. Partner with Corbis.com (database of 70 million images founded by Bill Gates) and Rolling Stone.com, which users can search separately from or in combination with AltaVista's own index of the Web. Consult the help files for more detailed information on performing an image search.
Google Image Search
"The Most Comprehensive Image Search on the Web"
Launched in June 2001, Google introduced its new image search engine. According to Google's always-useful FAQ page, Google "analyzes the text on the page adjacent to the image, the image caption and dozens of other factors to determine the image content. Google also uses sophisticated algorithms to remove duplicates and ensure that the highest quality images are presented first in your results." Since March 2001, the database of images has grown from 330 million to 880 million images! The Advanced Image Search feature provides methods to limit search queries by file size, file type, color, a specific Internet domain, etc.
General search engines often, however, provide inaccurate search results simply because they are not designed to evalulate the visual content of digital images. Currently, there is also a trend for many search engines to partner with large, proprietary image banks (e.g, Corbis, Inc.), making digital image research unnecessarily complicated for the average user.
Image Bots Promise To Be Even Better
An alternative to the idea that general search engines can provide one-stop-shopping for a user's Web needs is the content-based image retrieval (CBIR) application or multimedia web crawler. These "image bots" are no longer new to the Web, but offer different methods to improve the rudimentary search capabilities of most Internet search tools. Still being developed and refined, two examples of content-based visual retrieval technology include:
WebSEEk: Content-based Image & Video Search Tool for the Web
Developed at Columbia University. A collection of over 665,000 images which can be searched by category (animals, art, music, travel, etc.), free-text/keywords, and according to their visual content. Note: Site is frequently down or not working properly.
A proprietary, powerful knowledge discovery platform "that allows users to personalize search results through the use of dynamically generated classifications...regardless of the format, language or storage location of the retrieved data." The software can efficiently and accurately index graphics, illustrations, animation, and video. The search functions of commercial sites, such as eRugGallery.com, are also being powered by RetrievalWare®. Visit the website for eRugGallery.com and try the "Advanced Search" option to preview thousands of handmade rugs according to their shape, size, background color, border color, etc.