Multimedia Data Indexing

Multimedia (MM) data indexing refers to the problem of preprocessing a database of MM objects so that they can be efficiently searched for on the basis of their content. Due to the very nature of MM data, indexing solutions are needed to efficiently support similarity queries, where the similarity of two objects is usually defined by some expert of the domain and can vary depending on the specific application. Peculiar features of MM indexing are the intrinsic high-dimensional nature of the data to be organized, and the complexity of similarity criteria that are used to compare objects. Both aspects are therefore to be considered for designing efficient indexing solutions.