Events

Date	Topics	Speaker	Abstract	Host
Oct. 15, 2001	Vision-Based Hand Gesture Tracking and Recognition	Prof. Tomas S. Huang From: University of Illinois	This research is motivated by applications in human-computer interaction such as display control in virtual environments and the manipulation of virtual objects. Although we have studied both Appearance-Based and 3D Model-Based approaches, this talk will concentrate on the latter. Learn More	Steve Kang and Richard Hughey
Nov. 16, 2001	Bayesian Analysis of Image Sequences: Detection and Tracking of Complex Objects	Dr. David Fleet From: Xerox Palo Alto Research Center	Visual motion analysis concerns the estimation and recognition of motion from image sequences. Fundamental tasks include the estimation of image velocity (e.g., for video compression), the detection and estimation of scene structure (e.g. locating surface boundaries), and the detection and tracking of objects (e.g. 3D human motion capture from video). Learn More	Hai Tao
Nov. 19, 2001	Vision-Based Modeling and Interaction	Dr. Zhengyou Zhang From: Microsoft Research	Visual motion analysis concerns the estimation and recognition of motion from image sequences. Fundamental tasks include the estimation of image velocity (e.g., for video compression), the detection and estimation of scene structure (e.g. locating surface boundaries), and the detection and tracking of objects (e.g. 3D human motion capture from video). Learn More	Hai Tao
Nov. 29, 2001	Recent Work in Computer Imaging at FX Palo Alto Laboratory	Jonathan Foote, Don Kimber and Qiong Liu From: FX Palo Alto Laboratory	We present an overview of recent work at our laboratory, including FlyCam, a panoramic video imaging system; and FlyAbout, a virtual-reality system using spatially indexed panoramic video. We will give an overview of each system and (network permitting) a brief demo. Learn More	Hai Tao
Jan. 22, 2002	3-D Navigable TV System - Experiments on Algorithms and Architectures	George Chen From: ST Mico Electronics	This talk presents some experimental work towards building a prototype 3-D navigable TV system. The main focus is on the developement of algorithms and he investigation of suitable microprocessor architectures.Learn More	Hai Tao
Mar. 13, 2002	Recognizing Objects when the Lighting Changes	David Jabobs From: NEC Research Institute	Variations in lighting can have a significant impact on the appearance of an object. In this talk I will discuss a novel characterization of this variability for the case of Lambertian (non-shiny) objects. First, we represent lighting using spherical harmonics, and describe the effects of Lambertian materials as the analog of a convolution; this is similar to working in the frequency domain in signal processing. We are then able to show that almost all the appearance of Lambertian objects is determined by the first nine components of the lighting when represented as spherical harmonics. This allows us to represent the set of images an object produces with an analytically constructed 9D linear space, explaining prior empirical results. Learn More	Suresh Lodha
May 15, 2002	Shape Recognition for Automated Medical Diagnosis	Carlo Tomasi & Burak Gokturk From: Stanford University	This talk describe a novel framework for 3D shape representation and recognition. The main application is the automated diagnosis of colon cancer from Computerized Tomography (CT) scans. Cancer polyps are recognized through statistical classification of special geometric features computed via a new 3D pattern-processing algorithm, called the Random Orthogonal Shape Sections (ROSS) method. Learn More	Roberto Manduchi
Jul. 03, 2002	From Muybridge to Virtual Actors: The Mathematics of Motion Pictures	Chris Bregler From: Stanford University	The use of motion capture in science and entertainment includes new applications in computer animation, vision, bio/medical disciplines, and human computer interfaces. We survey these trends, touch on the underlying mathematics, and reveal the roots of today's motion pictures, seen over a century ago in Eadweard Muybridge's famous Motion Studies and the motion recordings at Disney's Animation Studios.Learn More	Roberto Manduchi
Jan. 15, 2003	Achieving Visual Ubiquity with Mobile or Widely-Distributed Sensors	Hector Gonzalez-Banos From: Honda Research Institute USA, Inc.	The goal of Ubiquitous Sensing is to measure or sense a physical property that exists distributed over multiple places at once. In this talk, I will describe a few of the combinatorial and motion planning issues that arise in ubiquitous sensing with image sensors. First, sensors are assumed to be stationary or almost stationary, so the only constraints present are geometric. This is the classic art gallery problem. If the sensors move and the workspace remains stationary the problem becomes the watchman route problem is based on the concept of illumination cones. For the second method, we introduce another affinity measure based on image gradient comparisons. Learn More	Hai Tao