My main research directions include pattern recognition, computer vision, machine learning, and image/signal processing. I have worked on some fundamental problems and contributed to the frontier knowledge in these areas. Some of the techniques have been used to conduct advanced multimedia applications. I am also interested in developing sparse-signal processing algorithms for both analysis and synthesis purposes. The techniques introduced are benefit to applications such as visual surveillance, human-computer interaction, and image/video analysis. Recent research topics include multiple-kernel learning, object detection and recognition, visual surveillance systems, facial image understanding, and deep learning.