Projects

	Learning Deep Representation with Large-scale Attributes
	DeepID-Net
	Joint Deep Learning for Pedestrian Detection
	Multi-task Recurrent Neural Network for Immediacy Prediction
...	A Discriminative Deep Model for Pedestrian Detection with Occlusion Handling
	Single-Pedestrian Detection Aided by 2-Pedestrian Detection
	Pedestrian Parsing via Deep Decompositional Neural Network
	Learning Mid-level Filters for Person Re-identification
	Visual Tracking with Fully Convolutional Networks
	Deeply Learned Attributes for Crowded Scene Understanding
	Scene-Independent Group Profiling in Crowd
	Cross-scene Crowd via Deep Convolutional Neural Networks
	Measuring Crowd Collectiveness
	Learning Collective Crowd Behaviors with Dynamic Pedestrian-Agents
	Correspondence-Free Multi-Camera Activity Analysis and Scene Modeling
	Trajectory Analysis and Semantic Region Modeling Using A Nonparametric Bayesian Model
	Unsupervised Activity Perception by Hierarchical Bayesian Model
	Learning Semantic Scene Models by Trajectory Analysis
	Discover Object Classes from a Collection of Images without Supervision Using the Spatial Dirichlet Allocation Model
	Deep Convolutional Network Cascade for Facial Point Detection
	Face Sketch Synthesis and Recognition
	Diffusion Tensor MRI Predictors of Cognitive Impairment in Confluent White Matter Lesion