Vision Problems from two face images? Stereo? Aging? Color from Gray? Face Recognition? Novel View Rendering? More???

Bio:

___________________________________________________________________________________________________________

I graduated with my PhD in Computer Science from Johns Hopkins University in May 2007, and my advisor is Prof. Gregory D. Hager. My thesis committee members are Professors Gregory Hager, Michael Kazhdan, Laurent Younes and Jianbo Shi (UPenn). Since October 2006, I have been working as a Research Scientist in Siemens Corporate Research, Inc., Princeton, New Jersey. I currently focus on solving medical imaging problems, such as topological structure parsing, anatomic structure detection and segmentation using general computer vision and applied machine learning techniques.

PhD in Computer Science, Johns Hopkins University, Baltimore, Maryland, May 2007

MSE in Computer Science, Johns Hopkins University, Baltimore, Maryland, May 2004

Advisors:
Prof. Gregory Hager, Johns Hopkins University, 2001-2007
Dr. Harry Shum, Microsoft Research Asia, 1999-2001
Prof. Zhanyi Hu, National Laboratory of Pattern Recognition, Institute of Automation, CAS, 1997-2000

Recent Talks:

_________________________________________________________________________________________________________

RSNA Talk on ICV detection for Colon CAD (Annual Meeting of Radiology Society of North America, Nov. 2007, Chicago)

ICML Talk on Clustering (International Conference of Machine Learning, June 2006, Pittsburgh), (ppt, zip)

CVPR Talk on Nonparametric Robust Visual Tracking (IEEE Conference on Computer Vision and Pattern Recognition, June 2007, Minneapolis), (ppt, zip)

I Gave a Tech Talk at Google in May 2006! The video was an overview of my graduate research on image patch representations/modeling for visual recognition and tracking. During the past year, some parts have been updated.

Research Interests:

__________________________________________________________________________________________________________

Structure Parsing for Medical Imaging and Analysis (Ongoing)

Supervised Learning Approach for Medical Imaging and Analysis (CVPR 2008, ECCV 2008, Ongoing)

Supervised Discriminative Dimension Reduction (Ongoing)

Online Appearance Modeling for Tracking and Recognition (NIPS 2006, CVPR 2007)
New Clustering Method for Visual Data (ICML 2006)
Scene Analysis and Category Recognition (CVPR 2005)
Articulated Object Motion Modeling and Recognition (NIPS 2004)
Efficient Particle Filtering using RANSAC (FPIV 2004, IVC 2006)   
Real-time Video Mosacing with Medical and Nonmedical Applications (CVPR 2003 Demo)
Graphical Models and Tree-Structured Object Tracking

Publications (refereed conference and journal):

___________________________________________________________________________________________________________

2008

Le Lu, Adrian Barbu, Matthias Wolf, Jianming Liang, Luca Bogoni, Marcos Salganicoff, and Dorin Comaniciu, "Simultaneous Detection and Registration for Ileo-Cecal Valve Detection in 3D CT Colonography", accepted in ECCV'2008: European Conference on Computer Vision, October, 2008, Marseille, France.

Le Lu, Adrian Barbu, Matthias Wolf, Jianming Liang, Marcos Salganicoff, and Dorin Comaniciu, "Accurate Polyp Segmentation for 3D CT Colonography Using Multi-Staged Probabilistic Binary Learning and Compositional Model", CVPR'2008: IEEE Conference on Computer Vision and Pattern Recognition, June, 2008, Anchorage, USA.

2007

Le Lu, "Image and Video Exploration by Classification and Clustering Using Global and Local Visual Features", PHD Thesis, Computer Science Department, Johns Hopkins University, Baltimore, Maryland, USA, April 2007.

Le Lu and Gregory D. Hager, "A Nonparametric Treatment on Location/Segmentation Based Visual Tracking", CVPR'2007: IEEE Conference on Computer Vision and Pattern Recognition (Oral), June, 2007, Minneapolis, USA.

2006

Le Lu and Gregory D. Hager, "Dynamic Background/Foreground Segmentation From Images and Videos using Random Patches", NIPS'2006: Neural Information Processing and System, Vancouver, B.C. Canada, Dec. 2006.

Le Lu and Rene Vidal, "Combined Central and Subspace Clustering on Computer Vision Applications", ICML'2006: International Conference of Machine Learning, June 2006, Pittsburgh, USA.

Le Lu, Xiangtian Dai and Gregory D. Hager, "Efficient Particle Filtering Using RANSAC with Application to 3D Face Tracking", Journal of Image and Vision Computing, vol 24, issue 6, pp.581-592, June 2006.

2005

Le Lu, Kentaro Toyama and Gregory D. Hager, "A Two Level Approach for Scene Recognition", CVPR'2005: IEEE Computer Society Conference on Computer Vision and Pattern Recognition, June 2005, San Diego, USA.

Two lessons learned: Lesson 1 (Image patch level Modeling): There is no way, in general, to distinguish the scene category properly for every given image patch without contextual information. Even with contextual constraints, it is still a very difficult problem if working with a large amount of "real" customer pictures, not professional photos like CorelDraw database. We try to argue that a mixture probability density function is trainable for image patches with stability in terms of modeling/discrimination. For a high level, like scene recognition, task, we can further learn the importance of each patch based material class in proportion to its confusion factor across different scene categories. Lesson 2 (Both levels): Given a complex, high dimensional distribution of data samples, try to learn a mixture density model (considering smoothness offered by a density function) in an efficient and discriminative way f needed. For a recognition/classification task, try to integrate discriminative information within the overall generative (density) framework.

2004

Le Lu, Gregory D. Hager and Laurent Younes, "A Three Tiered Approach for Articulated Object Action Modeling and Recognition", NIPS'2004: Neural Information Processing and System, Vancouver, B.C. Canada, Dec. 2004. (a version with slightly more details)

Le Lu, Xiangtian Dai and Gregory D. Hager, "A Particle Filter without Dynamics for Robust 3D Face Tracking", IEEE Workshop of Face Processing in Video jointed with CVPR'2004, June 2004, Washington DC, USA.

Long Quan , Yichen Wei , Le Lu and Heung-Yeung Shum, "Constrained planar motion analysis by decomposition",  Journal of Image and Vision Computing, vol 22, issue 5, pp. 379-389, May 2004. 

2003

Xiangtian Dai, Le Lu and Gregory D. Hager,  "Real-time Video Mosaicing with Adaptive Parameterized Warping", Demo Program, CVPR'2003: IEEE Computer Society Conference on Computer Vision and Pattern Recognition, May 2003, Midson, USA.

2002

Le Lu and Hung-Tat Tsui, "Algebraic Characteristic, Geometric Interpretation and Degeneracy Analysis of Planar Motions, with Applications to Camera Self-calibration ", extended version submitted to Journal of Pattern Recognition.

Le Lu and Hung-Tat Tsui, "Algebraic Characteristic and Geometric Interpretation of Planar Motions and their Applications to Camera Self-calibration",  ACCV'2002: the 5th Asian Conference on Computer Vision, Jan. 2002, Melbourne, Australia (initially invited for a special issue of International Journal of Image and Graphics for ACCV 2002, 10 out of all submissions)

2001

Le Lu, Zhengyou Zhang, Heung-Yeung Shum, Zhicheng Liu and Hong Chen, Model- and Exemplar-based Robust Head Pose Tracking Under Occlusion and Varying Expression, IEEE Workshop on Models versus Exemplars in computer Vision jointed with CVPR'2001, Dec. 2001, Hawaii, USA. 

Long Quan, Le Lu, Heung-Yeung Shum and M. Lhuillier, Concentric Mosaic(s), Planar Motion and 1D Cameras, ICCV'2001: The 8th international Conference on Computer Vision, pp.193-200, Jul. 2001, Vancouver, Canada.

Le Lu and Zhanyi Hu, "A New Factorization Technique for Projective Reconstruction", Chinese Journal of Electronics, (English Version), No.2, pp.196-202, Feb 2001.

2000

Le Lu, Hung-Tat Tsui and Zhanyi Hu, "A Novel Planar Motion Detection Method and the Robust Estimation of 1D Trifocal Tensor", ICPR'2000: The 15th International Conference on Pattern Recognition, Vol. 3, pp.815-818, Sept. 2000, Barcelona, Spain. (ACM Library link)

Le Lu, Zhanyi Hu and Hung-Tat Tsui, "Sub-sequence Factorization - an Efficient Approach for Projective Reconstruction", ACCV'2000: The 4th Asian Conference on Computer Vision, pp.1052-1057, Jan. 2000, Taipei, Taiwan, China. 

Publications (Technical Report):

Le Lu, "Real-time Video Mosaicing with Adaptive Parameterized Warping", Computational Interaction and Robotics Lab, Technical report, Computer Science Department, Johns Hopkins University, 2003.  (example and code)

Affiliations:

_______________________________________________________________________________________________________

Integrated Data Systems Dept., Siemens Corporate Research, Inc. 10/2006-now

Computer Science Dept., Johns Hopkins University, 08/2001-09/2006

Interactive Visual Media Group, Microsoft Research, Redmond, (Intern), 06/2004-08/2004

Visual Computing Group, Microsoft Research, Beijing, (Intern), 12/1999-07/2001

Electronic Engineering Dept., Chinese University of Hong Kong (Visiting student), 03/1999-08/1999

National Laboratory of Pattern Recognition, Institute of Automation, CAS, 09/1996-12/1999

Mentors (collaborators):

_______________________________________________________________________________________________________
Prof. Laurent Younes, Center of Imaging Science, Johns Hopkins university
Prof. Rene Vidal, Center of Imaging Science, Johns Hopkins University
Dr. Kentaro Toyama, Microsoft Research India, summer 2004
Prof. Long Quan, HKUST, 2000
Dr. Zhengyou Zhang, Microsoft Research, 2000-2001

Contacts:

Siemens Corporate Research, Inc.

755 College Road East, Princeton, NJ 08540

Tel: 610-219-4369

Email: lelu@cs.jhu.edu