Visual Representation Bridging Machine Learning to Computer Vision

Stay connected



Share on facebook
Share on twitter
Share on linkedin

CIS Colloquium, Oct 29, 2008, 10:30AM – 11:30AM, Wachman 447

Visual Representation Bridging Machine Learning to Computer Vision

Le Lu, Siemens Research

In this talk, I will first emphasize on the global view of finding the “right” visual representations which brings effective machine learning techniques to meet and tackle a variety of computer vision applications, given a problem solving condition. The example applications include 2D/3D location, pose and segmentation based tracking, action recognition, scene recognition, clustering on visual data sets and 3d medical imaging tasks of anatomic structure detection and segmentation. In the second part, as a special case study, I will walk through an image patch based density representation to learn scene categories and a nonparametric image patch based dynamic appearance model for location and segmentation driven visual tracking tasks.

Sampled publications are available at

Le Lu received his Ph.D. and MSE degrees of Computer Science from Johns Hopkins University in 2007 and 2004 respectively, focusing on computer vision and statistical learning (advised by Prof. Gregory Hager). He spent three years in National Laboratory of Pattern Recognition, Institute of Automation, Chinese Academy of Sciences from 1996 to 1999, studying computational geometry based 3D computer vision. He also received his B.E. in Precision Instrument and Automatic Control from the Beijing Polytechnic Univerisity, China, in 1996.

Since October 2006, Le Lu has been a Research Scientist in Integrated Data Systems Department, Siemens Corporate Research, at Princeton, New Jersey. He also currently holds a staff scientist position at Knowledge and CAD solutions, Siemens Medical Solutions, at Malvern, Pennsylvania. During the summer of 2004, he was an Intern with Dr. Kentaro Toyama, at Interactive Visual Media Group, Microsoft Research (Redmond). He also interned with Dr. Harry Shum at Visual Computing Group, Microsoft Research (Beijing) from Dec. 1999 to July 2001. From March to August in 1999, he was a visiting graduate student in Department of Electronic Engineering, Chinese University of Hong Kong, Hong Kong, China.