Today’s World Wide Web is truly a video Web. Millions of video clips are available to users instantly thanks to widely available broadband IP networks, low-cost storage and mature digital video delivery technologies. The content of this video runs the gamut from skateboarding antics captured on mobile phone cameras up through graduate level university lecture series on computer science. On the commercial entertainment side, all major broadcasters and movie studios have on-line strategies which range from a focus on promoting traditional distribution channels through releasing primetime programming through the Web to capture an emerging demographic who increasingly turn to their laptops for video entertainment instead of their televisions.
Although Internet video systems have made great strides, television still provides the highest quality digital video available to consumers on a daily basis at a level of quality far beyond that of traditional best-effort Internet video streaming. On-line high definition (HD) content is still a novelty. Cable, direct broadcast satellite, and over-the-air digital broadcast are mature technologies providing HD quality entertainment to millions of consumers today, and Internet protocol television (IPTV) is emerging to provide more functionality and provide increased convergence with existing Web technologies. High capacity digital video recorders allow consumers to easily capture many hours of content for viewing at their convenience. An increasing array of set-top devices and smart TVs with IP connectivity provides access to the wealth of Web video (e.g. expatriates can view news from home in their native langue) via streaming or on demand. Closer to home, consumers can browse their videos and photos captured from their digital cameras or purchased on line and archived on their home network server.
The video Web extends beyond fixed appliances to mobile devices which support fully functional browsers and media players. Battery life and user consumption contexts may limit viewing of long-form content, but specially designed short-form content provides valuable entertainment for opportunistic consumption scenarios such as waiting for a late bus.
This dazzling array of options for access to high quality video anytime, anywhere provides opportunities for service and technology providers who can develop services and tools to help users manage their media experience and locate content of interest from the vast ocean of irrelevant or even repulsive material. Service providers plan to build universal three-screen services to allow users to seamlessly switch from TV to PC to mobile viewing. In fact, three screens are not enough to encompass all of today’s consumption scenarios and the term “any screen” is used to include personal media players, portable gaming devices and others such as Internet connected picture frames.
Given the potential impact of technological breakthroughs in video services, it is not surprising that there is no shortage of academic and industrial research groups focused on this task. Further, successful solutions require an interdisciplinary approach, drawing from diverse fields including information retrieval, natural language processing, data mining, machine learning, multimedia databases, as well as speech, audio, image and video processing [Haupt05]. Data visualization, user interface design, human computer interfaces and the consideration of social aspects of media consumption and interaction such as rating, tagging and recommendation are of equal or perhaps greater importance than the media processing technologies. Improvements to the start of the art of video search can draw from a broad base indeed.
