Intelligent signal processing for future digital media systems

点击翻译按钮获取中文摘要

The objective of this proposal, entitled `Intelligent signal processing for future DIgital MEdia Systems', (I-DIMES) is to research promising ideas in the design, development and application of intelligent signal processing solutions that provide cognitive processing of digital content, model and analyze complex human behaviour, and facilitate autonomous recognition of context in a process-centric digital environment. This program attempts to provide answers to three interlaced problems, namely: (a) attention problem - how to extract the relevant digital information; (b) interface problem - what is the best way to present the information when the conditions have changed; and (c) application problem - what are the best uses for the extracted information. Thus, the long term research objectives of this proposal are to: (i) provide a comprehensive and systematic study of the links between digital content, context learning, situation awareness, activity recognition, and adaptation; (ii) devise a generalized processing framework for activity and behaviour recognition as part of a digital media environment. The short term objectives are to: (i) personalize access to visual content using end-users' preferences and profiles; (ii) organize content in order to enhance processing, retrieval, coding, and transmission of visual content; (iii) empower users through content and context fusion; (iv) introduce an integrity preservation and privacy enhancing framework; and (v) exploit novel applications of the developed solutions. In addition to having significant theoretical implications, the proposed research program has potential for social impact through the development of technological solutions for enhancing the life of individuals. To this end, an application scenario has been chosen to test the applicability of the proposed methodologies. The scenario comes from the field of EEG-based brain-computer interfaces and it will be used to test user-intention inference and cognitive processing recognition procedures.