Streaming-Media Knowledge Discovery

T he proliferation of powerful Web-based search engines that efficiently index and search large document collections has made it easier for the casual user to search and browse text documents on the Web. Although several portals provide comprehensive text information, we have yet to see a similarly comprehensive resource list for locating relevant information in streaming media, whether via search engines, Web directories, or portals. Yet the amount of Web-based streaming media grows continually. For example, a recent study estimates that live media channels on the Web number in the thousands. In addition, advances in storage, computing power, and bandwidth have resulted in increasing deployment of streaming video in applications such as workplace training, distance education, entertainment, and news. The Multimedia Knowledge Discovery group at IBM Almaden Research Center explores technologies that help locate streaming media on the Web, build effective indexing and classification tools for streaming media, and develop compelling streaming-media applications. The “What Is Streaming Media?” sidebar explores this emerging field in detail. Our ongoing work, motivated by the Web’s role as an ever-increasing information source and knowledge repository for many applications, focuses on locating, analyzing, and indexing Web-based streaming media. Frequently, knowledge management applications and information portals synthesize unstructured text information from the Web, intranets, and partner sites. Given this context, we crawl a statistically significant number of Web pages, detect those that contain streaming media links, crawl the media links to extract associated metadata, then use the crawl data to build a resource list for Web media. We have used these crawl-data findings to build a media indexing application that uses content-based indexing methods.