Multimedia Retrieval

Finding multimedia content calls for different methods than querying conventional structured databases. Searching is the central difficulty in multimedia databases, and researchers have approached it from many directions, including information retrieval, human-computer interaction, computer vision and pattern recognition, well beyond the usual database field.

Drawing on over a decade of classroom experience, Henk M. Blanken and his coeditors have gathered the material normally taught in advanced undergraduate and graduate courses on multimedia retrieval and multimedia databases. The chapters cover the overall architecture of multimedia retrieval systems and metadata languages such as Dublin Core, RDF and MPEG. They also address pattern recognition with Markov models, unsupervised learning and clustering, indexing methods for audio and video streams, interaction and control, the protection of content and user privacy, and the measurement of search effectiveness and efficiency.

Throughout, the authors focus on high-level features and show how such features feed into mathematical models that drive the retrieval process. Each chapter ends with suggestions for further reading, and additional exercises and teaching materials are available online.