Currently I am an Assistant Research Scientist in Computer Vision Lab, Institue for Advanced Computer Studies(UMIACS) at University of Maryland at College Park, working with Prof. Larry S. Davis. My research interest includes computer vision, pattern recognition and machine learning, specifically on action detection/recognition, object detection/tracking/categorization, sparse coding and dictionary learning.

[Publications | Research Projects | Links | My CV]

Selected Publications:

Conference:

  • Qiang Qiu, Zhuolin Jiang, Rama Chellappa. "Sparse Dictionary-based Representation and Recognition of Action Attributes". IEEE Conference on Computer Vision, 2011. [pdf]
  • Zhuolin Jiang, Zhe Lin, Larry S. Davis. " Learning a Discriminative Dictionary for Sparse Coding via Label Consistent K-SVD". IEEE Conference on Computer Vision and Pattern Recognition, 2011. [pdf][slide][poster]
  • Zhuolin Jiang, Zhe Lin, Larry S. Davis. "A Tree-based Approach to Integrated Action Localization, Recognition and Segmentation". Third Workshop on Human Motion (In Conjuntion with ECCV), 2010. [pdf][slide]
  • Zhe Lin, Zhuolin Jiang, Larry S. Davis. "Recognizing Actions by Shape-Motion Prototype Trees". IEEE Conference on Computer Vision, 2009. [pdf] [slide][project page]

Journal:

  • Zhuolin Jiang, Zhe Lin, and Larry S. Davis, "Recognizing Human Actions by Learning and Matching Shape-Motion Prototype Trees". IEEE Transactions on Pattern Analysis and Machine Intelligence, 2012, 34(3): 533-547. [pdf]
  • Zhuolin Jiang, Zhe Lin, and Larry S. Davis, "Class Consistent k-means: Application to Face and Action Recognition". Computer Vision and Image Understanding, 2011. (under major revision).

Research Projects:

Sparse Dictionary-based Representation and Recognition of Action Attributes

We present an approach for dictionary learning of action attributes via information maximization. We unify the class distribution and appearance information into an objective function for learning action attributes. The objective function maximizes the mutual information between what has been learned and what remains to be learned in terms of appearance information and class distribution for each dictionary item. We propose a Gaussian Process (GP) model for sparse representation to optimize the dictionary objective function. The sparse coding property allows to realize a very efficient dictionary learning process.

Learning a Discriminative Dictionary for Sparse Coding via Label Consistent K-SVD (LC-KSVD)

We propose a label consistent K-SVD algorithm to learn a discriminative dictionary for sparse coding. We associate label information with each dictionary item to enforce discriminability in sparse codes during the dictionary learning process. More specifically, we introduce a new label consistent constraint called "discriminative sparse-code error" and combine it with the reconstruction error and the classification error to form a unified objective function. The optimal solution is efficiently obtained using the K-SVD algorithm. It yields dictionaries so that feature points with the same class labels have similar sparse codes.

A Tree-based Approach to Integrated Action Localization, Recognition and Segmentation

We propose a tree-based approach to integrated action segmentation, localization and recognition. A binary tree model is constructed using the set of learned action prototypes. In this tree model, each leaf node corresponds to a prototype and contains a list of parameters including the probability of the prototype belonging to an action category, frame indices of all the training descriptors which matched to this prototype, and a rejection threshold. These parameters allow us to integrate action localization, recognition and segmentation in a video.

Recognizing Human Actions by Shape-Motion Prototype Trees

We propose a prototype-based approach for action recognition. Our approach represents an action as a sequence of prototypes for efficient and flexible action matching in long video sequences.Our approach captures correlations between shape and motion by learning action prototypes in the joint feature space, but performs recognition efficiently via tree-based prototype matching and look-up table indexing. In addition, it has the advantage of tolerating complex dynamic backgrounds due to median-based background motion compensation and probabilistic frame-to-prototype matching.

Important Links:

Related Links for CV:

Important Dataset:
Latest update 12-06-2011