The goal of this book as posed in Chapter 1 is to investigate what techniques are necessary to be added to a database management system to support content-based video retrieval. The research that has been done to accomplish this goal is positioned at the crossroads among many fields that include among others databases, information retrieval, computer vision, image processing, and pattern recognition. All these research communities have identified the problem of bridging the semantic gap as the major problem in content-based video retrieval. In this book, we have demonstrated that the semantic gap can be closed in restricted domains by the integrated use of the spatio-tempoal approach, HMMs, and DBNs. Another major contribution of the work presented in the book is the integration of these technologies within the DBMS, which allows the system to be used in multiple domains with minimal adjustments and provides a user with dynamic feature and semantics extraction.
KeywordsHide Markov Model Bayesian Network Video Content Relevance Feedback Video Processing
Unable to display preview. Download preview PDF.