Image and Video Indexing and Retrieval

Abstract

This chapter explores the latest technologies in image and video retrieval. It describes several methods for extracting features that are used to measure image and video similarity in multimedia databases. It also describes techniques to bridge the gap between low-level features and high-level semantics. A typical content-based image/video retrieval system includes three major aspects-feature extraction, high-dimensional indexing, and system design. Among the three aspects, high-dimensional indexing is important for speed performance; system design is critical for appearance performance; and feature extraction is the key to accuracy performance. The accuracy performance of a retrieval system is very subjective and user dependent. To a user, the similarity among objects is often high level or semantic. However, features extracted from objects are often low-level features as most of them are extracted directly from digital representations of objects in the database. The gap between low-level features and high-level semantics has been a major obstacle to a better retrieval performance.