Shanghang Zhang

Propose the topology adaptive graph convolutional network, a novel graph convolutional network that generalizes CNN architectures to graph-structured data and provides a systematic way to design a set of fixed-size learnable filters to perform convolutions on graphs. The topologies of these filters are adaptive to the topology of the graph when they scan the graph to perform convolution, replacing the square filter for the grid-structured data in traditional CNNs. It can be used with both… Show more

Propose the topology adaptive graph convolutional network, a novel graph convolutional network that generalizes CNN architectures to graph-structured data and provides a systematic way to design a set of fixed-size learnable filters to perform convolutions on graphs. The topologies of these filters are adaptive to the topology of the graph when they scan the graph to perform convolution, replacing the square filter for the grid-structured data in traditional CNNs. It can be used with both directed and undirected graphs. Show less

• Generate spatially-variant blur responses using fully convolutional neural networks.
• Understand if such responses are undesired by distilling higher-level image semantics: Learn salient object segmentation map and content feature map to localize important content in the images.

• Propose a new generalization bound for domain adaptation when there are multiple source domains with labeled instances and one target domain with unlabeled instances.
• Propose an efficient implementation of the theoretical results using adversarial neural networks: Learn feature representations that are invariant to the multiple domain shifts while still being discriminative for the learning task.

Develop a multi-task learning scheme based on neural networks for robot action prediction, which is very important step for autonomous robots design.

Develop a Convolutional Variational Auto-Encoder to generate features of percepted images for the robot action prediction, which is capable of capturing the useful statistics of robot actions without requiring large-scale training samples or hand-engineered features.

I develop a deep multi-task model to jointly estimate vehicle density, segment foreground and detect vehicles based on fully convolution networks to overcome the challenges of web camera data, such as low resolution, high occlusion, and large perspective. Multi-domain adaptation mechanism is explored to adapt the deep model to different cameras and environmental conditions. Filters in each convolution layer are dynamically generated to learn different camera perspectives. Deep spatio-temporal… Show more

I develop a deep multi-task model to jointly estimate vehicle density, segment foreground and detect vehicles based on fully convolution networks to overcome the challenges of web camera data, such as low resolution, high occlusion, and large perspective. Multi-domain adaptation mechanism is explored to adapt the deep model to different cameras and environmental conditions. Filters in each convolution layer are dynamically generated to learn different camera perspectives. Deep spatio-temporal networks are developed to incorporate the temporal information of traffic flow. Show less

FCN based multi-task learning to jointly estimate vehicle density and vehicle count with end-to-end dense (pixel-level) prediction. It allows arbitrary input image resolution, and adapts to different vehicle scale and perspective.

Deep spatio-temporal networks are developed to incorporate the temporal information of traffic flow.

Optimization based vehicle density estimation with rank constraint to embed road geometry in the weight matrix and significantly reduce error induced by… Show more

FCN based multi-task learning to jointly estimate vehicle density and vehicle count with end-to-end dense (pixel-level) prediction. It allows arbitrary input image resolution, and adapts to different vehicle scale and perspective.

Deep spatio-temporal networks are developed to incorporate the temporal information of traffic flow.

Optimization based vehicle density estimation with rank constraint to embed road geometry in the weight matrix and significantly reduce error induced by camera perspective.

Collect and annotate a large-scale webcam traffic dataset, which poses new challenges to traffic density estimation algorithms and other learning based methods. It is the first and largest webcam traffic dataset with elaborate annotations. Show less

Develop DCNN to recognize products in the query image and discover complementary views on the brand associations by mining
large-scale online images.

Design deep convolutional neural network (DCNN) to detect and understand obstacles around driving vehicles.

Hierarchical feature extractor to adapt the network particularly to autonomous driving without overfitting.

Finalist Awards for Qualcomm Innovation Fellowship (35 outside 146 teams from Engineering Top 10 Univerisities in U.S.)

Layout images clustering to discover latent fail patterns. Clustering images of the layout locations to discover the common layout features and the latent fail patterns.

Spatial variations prediction based on Bayesian model fusion. Propose Bayesian model fusion (BMF) to model the spatial correlation and predict spatial variations on a chip.

Working out a novel surveillance video compression scheme based on the background modeling and foreground segmentation.

Key Algorithms Research on Multimedia Processing and SoC Design.

In charge of high definition(HD) video codec and communication such as network bandwidth adaptation.