Multimedia content is nowadays abundant both in proprietary collections and on the web. Intelligent, targeted and seamless access to this content is a key requirement for a wide range of applications in the business, education, entertainment, cultural and other sectors. We have worked and have expertise on automatic audio-visual information analysis and understanding for intelligent access to multimedia content, including image segmentation and semantic analysis, video temporal segmentation to shots and scenes using multi-modal cues, spatio-temporal video segmentation, motion statistical processing and representation, activity detection and human action recognition, semantic shot classification, cross-media analysis and temporal context evaluation and high-level feature extraction from video. We also have expertise on ontologies and reasoning for multimedia representation and semantic analysis. We have developed image and video search engines both for generic content and for specific applications such as cultural image search, and have expertise on multimedia retrieval using content-based, context-based and semantic techniques; personalization and recommendation techniques for access to multimedia content; user interfaces for multimedia search and retrieval; and annotation tools for multimedia annotation and semi-automatic generation of indexing metadata.