I am currently a computer vision researcher at Sighthound Inc. I graduated from Dr. Mubarak Shah's computer vision lab at University of Central Florida. My research interests include computer vision, machine learning and computational photography.
M.S., Electrical Engineering, Shanghai Jiao Tong University, China, 2009
B.S., Electrical Engineering, Huazhong University of Science and Technology, China, 2006
Guang Shu, Afshin Dehghan, Mubarak Shah, "Improving an Object Detector and Extracting Regions using Superpixels", in CVPR, 2013. (PDF)
Yang Yang, Guang Shu, Mubarak Shah, "Semi-supervised Learning of Feature Hierarchies for Object Detection in a Video", in CVPR, 2013. (PDF)
Guang Shu, Afshin Dehghan, Omar Oreifej, Emily Hand, Mubarak Shah, "Part-based Multiple-Person Tracking with Partial Occlusion Handling", in CVPR, 2012. (PDF)
Omar Oreifej, Guang Shu, Teresa Pace, and Mubarak Shah, "A Two-Stage Reconstruction Approach for Seeing Through Water", in CVPR, 2011. (PDF)
Guang Shu, Lixiu Yao, xiaochao Yang, Xin Zuo, Jie Yang, "3D Facial Caricaturing Based on Sparse Morphable Model", in Acta Electronica Sinica, August 2010
Guang Shu, Lixiu Yao, Lu Chen, Jie Yang, "3D Face Reconstruction from one Image Based on a Generic Model", in Journal of Shanghai Jiao Tong University, March 2009
Xiaochao Yang, Yue Zhou, Guang Shu, Tianhao Zhang, "Manifold Learning based Gait Recognition using Gabor phase spectrum", in Acta Electronica Sinica, April 2009.
Xiaochao Yang, Yue Zhou, Tianhao Zhang, Guang Shu, Jie Yang, "Gait Recognition Based on Dynamic Region Analysis", in Signal Processing, September 2008.
Object Detection in Video using Superpixels (CVPR 2013)
Part-based Multiple-Person Tracking (CVPR 2012)
Proposed a multiple-person tracking method for highly crowded scenes.
Seeing through Water (CVPR 2011)
Introduced a sparse representation approach for removing distortions of underwater video.
Semantic Concept Detection (2013)
Developed a semantic concept detection system for TRECVID MED Challenge 2013. SRI-Sarnoff-UCF team won the second and ﬁrst place in TRECVID MED 2013-10EK and 0EK respectively.
NONA Tracking System (2010)
Developed a multi-threaded real-time tracking system for high-resolution videos using C++ and OpenCV.
Panoramic Video Summary (2010)
Developed a system to generate a panoramic video summary of detected actions of interest from a single pan-tilt-zoom camera.
3D Face Modeling and Caricaturing (2009)
Proposed an algorithm to reconstruct 3D face model from only one frontal image.
Proposed an algorithm to make 3D face caricatures.
Developed a 3D face modeling software using C++, OpenCV and OpenGL.