I’m a 5th year PhD student at UMass-Amherst. I am co-advised by Prof. Erik Learned-Miller and Prof. Subhransu Maji at the UMass Vision Lab. I got my master’s degree in Compute Science from Brown University and my bachelor’s degree in Intelligent Science and Technology from Peking University back in China.
My major research interests lie in computer vision and graphics. Below you can find some of my research projects.
A network architecture that efficiently operates on a sparse set of samples in a high-dimensional lattice.
Hang Su, Varun Jampani, Deqing Sun, Subhransu Maji, Evangelos Kalogerakis, Ming-Hsuan Yang, and Jan Kautz, "SPLATNet: Sparse Lattice Networks for Point Cloud Processing", CVPR 2018 (oral, to appear).
A novel CNN architecture that combines information from multiple views of a 3D shape into a single and compact shape descriptor offering state-of-the-art performance in a range of recognition tasks.
Ranked #1 in a SHREC'16 contest!
Hang Su, Subhransu Maji, Evangelos Kalogerakis, and Erik Learned-Miller, "Multi-view Convolutional Neural Networks for 3D Shape Recognition", ICCV 2015.
M. Savva, F. Yu, H. Su, M. Aono, B. Chen, D. Cohen-Or, W. Deng, H. Su, S. Bai, X. Bai, N. Fish, J. Han, E. Kalogerakis, E. G. Learned-Miller, Y. Li, M. Liao, S. Maji, A. Tatsuma, Y. Wang, N. Zhang, and Z. Zhou, "SHREC’16 Track: Large-Scale 3D Shape Retrieval from ShapeNet Core55", Eurographics Workshop on 3D Object Retrieval, J. Jorge and M. Lin, editors, 2016.
The first large-scale scene attribute database.
G. Patterson, C. Xu, H. Su, J. Hays, "The SUN Attribute Database: Beyond Categories for Deeper Scene Understanding", IJCV, May 2014.
In this project, I implemented in C++ a human face and body detection system based on the paper "Face detection, pose estimation and landmark localization in the wild" (X. Zhu and D. Ramanan, CVPR 2012). This C++ implementation achieved 0.95 recall and 0.90 precision on eHarmony’s user proﬁle photos.
The goal of this work is to automatically distinguish high quality professional photos from low quality snapshots.
Here I focus on assessing the quality of photos that contain faces (e.g. user profile photos). I propose several features specially useful for this task, e.g. skin smoothness, composition, bokeh, etc. Experiments show that with some small modifications they are also very useful for assessing other types of photos.
Onboard vehicle detection plays a key role in collision prevention and autonomous driving. Camera-based detection techniques have been proven effective and economical, and show extensive application prospect.
This project is focused on the development of a front vehicle detection system using onboard camera. Hypothesis generation based on shadows and hypothesis verification based on HOG features are integrated to achieve a real-time system. A passing vehicle detection component using optical flow is also proposed, and obtains fast and reliable detections.