Similarity search through fuzzy relational algebra

Multimedia informations systems and Web-based applications such as search engines need to automatically extract and classify multimedia data such as images and textual documents. For this reason, very significant research has focused on similarity queries from the standpoint of the definition and creation of appropriate index structures. We provide a formal framework in which formulate powerful similarity-based queries. For this purposes a fuzzy relational algebra extending the classical relational algebra is defined and equivalence and containment rules for query processing are given.