Visual computing is a combination of computer vision, machine learning, pattern recognition and other related fields. Due to the importance of ubiquitous visual data and promising applications, large-scale image / video processing and understanding have attracted strong interest of researchers. Although in recent years there have been some important progress, many challenging research issues remain unresolved, e.g., the limited description capacity of the existing visual features; in terms of visual representation, large-scale multi-source heterogeneous data lack effective analysis tools; the traditional classifier for large-scale, multi-class high-dimensional data is inefficient in visual learning and reasoning. For a major demand of China's information industry for advanced visual computing technologies, and for international cutting-edge research directions in the vision field, this project aims at the joint of outstanding vision researchers from China and Australia to study large-scale visual computing theories and methods, to breakthrough a major bottleneck in the efficiency, accuracy, robustness in terms of current visual computing technologies, and to develop visual feature extraction, representation, learning and reasoning methods for large-scale datasets, while providing the theoretical basis and key technical support for a wide range of promising applications such as wide-area video monitoring, multi-modal biometrics, and Internet mass visual information security.
(Sources of Funding: NSFC)