Among countless others, applications of multimedia databases include personal and public photo/media collections, personal information management systems, digital libraries, online and print advertising, digital entertainment, communications, long-distance collaborative systems, surveillance, security and alert detection, military, environmental monitoring, ambient and ubiquitous systems that provide real-time personalized services to humans, accessibility services to blind and elderly people, rehabilitation of patients through visual and haptic feedback, and interactive performing arts. This diverse spectrum of media-rich applications imposes stringent requirements on the underlying media data management layer. Although most of the existing work in multimedia data management focuses on content-based and object-based query processing, future directions in multimedia querying will also involve understanding how media objects affect users and how they fit into users’ experiences in the real world. These require better understanding of underlying perceptive and cognitive processes in human media processing. Ambient media-rich systems that collect diverse media from environmentally embedded sensors necessitate novel methods for continuous and distributed media processing and fusion schemes. Intelligent schemes for choosing the right objects to process at the right time are needed to allow media processing workflows to be scaled to the immense influx of real-time media data. In a similar manner, collaborative-filtering-based query processing schemes that can help overcome the semantic gap between media and users’ experiences will help the multimedia databases scale to Internet-scale media indexing and querying.

HETEROGENEITY

Most media-intensive applications, such as digital libraries, sensor networks, bioinformatics, and e-business applications, require effective and efficient data management systems.