This book offers a systematic introduction to an understanding-oriented approach to multimedia content analysis. It integrates the visual understanding and learning models into a unified framework, within which the visual understanding guides the model learning while the learned models improve the visual understanding. More specifically, it discusses multimedia content representations and analysis including feature selection, feature extraction, image tagging, user-oriented tag recommendation and understanding-oriented multimedia applications. The book was nominated by the University of Chinese Academy of Sciences and China Computer Federation as an outstanding PhD thesis. By providing the fundamental technologies and state-of-the-art methods, it is a valuable resource for graduate students and researchers working in the field computer vision and machine learning.
