Chinese Version

 

Xianghua Ying

Associate Professor
National Laboratory on Machine Perception
School of Electronic Engineering and Computer Science
Peking University
 
Contact Information:
National Laboratory on Machine Perception
Peking University
Beijing 100871, China 
Tel: (86-10)6275-9076
Fax: (86-10)6275-5569
E-mail: xhying@cis.pku.edu.cn

Xianghua Ying , male, born in 1976. He received PhD in Pattern Recognition and Intelligent Systems, from the Institute of Automation, Chinese Academy of Sciences, in 2004. Currently, he is an associate professor in School of Electronic Engineering and Computer Science, Peking University.

Office hours: Tuesday and Wednesday 16:00-17:30.

For who wants to talk to me: To reach me you can use e-mail (keeping your messages as short as possible) or phone. You can come in my office as well. You need to look at my time-table to make sure that I am here and you'd better come during my office hours: I will not kick you out if you come in another time but I can be absent or really busy.

Events Announcements

OMNIVIS 2005, the 6th Workshop on Omnidirectional Vision, Camera Networks and Non-classical Cameras, will be held on October 21st, 2005, in Beijing, China, in conjunction with ICCV 2005.

Research Interests

3D Reconstruction, Human Machine Interaction, Motion Analysis, Virtual Reality, Image-Based Modeling and Rendering, Camera Calibration, Omnidirectional Vision, etc.

 

Visual Compass: A Novel Virtual Reality 3D Rotation Input Device Using a Camera Pose Estimation Technique from a Single View of Parallel Lines

In many intelligent environments such smart rooms, wearable computing and tele-conferencing, conventional mice and keyboards seem not suitable, some intuitive, immersive and cost-efficient interaction devices are developed. There are too many implemented application systems, and we do not intend to cite them all. Here, we only introduce a representative vision-based interface system proposed by Zhang, et al (2001), called Visual Panel, which employs an arbitrary quadrangle-shaped panel and a tip pointer as an intuitive input device. The technique proposed in this paper can be employed to recover 3D rotation only from a set of parallel lines,therefore, we can use a rectangular parallelepiped, such as a packaging box of a bottle, which can be occluded by hands or other objects very seriously, just left three parallel lines visible which are sufficient to drive a virtual 3D objects rotated in the screen as the box do. Obviously, if there is no occlusion, it is easy to recover the 3D rotation parameters of the box. Though the packaging box is occluded by hands very seriously, but 3D rotation can yet be recovered. Therefore, this technique can be used in, such as, a virtual object exhibition system for persons who are not so familiar with the computer operations.

 

Real-time Camera Calibration from Image Sequences of Balls in Sport Scenes

Spherical objects and vanish points are often used for camera calibration. An occluding contour of a sphere is projected to a conic in the perspective image, and using a moving active camera, the trajectory of a vanishing point in the perspective images is also a conic when the camera is rotated about a fixed 3D axis whereas the translation of the camera is arbitrary. In fact, the problems of camera calibration using conics from spheres or vanishing points can be described by same mathematic representations. Two linear approaches to the problems are proposed in this paper: one based on the geometric interpretation of the relation between image conics and the image of the absolute conic, and the other using the special structure of the problems in algebra. Only three such conics are needed for the two linear approaches, and the minimum number for previous nonlinear optimization methods is also three. All five intrinsic parameters are recovered linearly without making assumptions, such as, zero-skew or unitary aspect ratio which are often used in previous methods. The two linear algorithms have been tested in extensive experiments with respect to noise sensitivity and also made comparisons with recent calibration techniques.

 

 

Walkthrough in Large Environments Using Concatenated Panoramas

In recent panorama-based virtual reality systems, there exists a key issue that is how to achieve continuous walkthrough between the two neighboring panoramas. In order to solve this issue, we firstly take images in the sampling points of the transition path, and then the coefficients of image transformations among these images are estimated directly from these images. Finally, the images in the non-sampling points are synthesized from those images in the sampling points using image transformation method. The architecture and schedule mechanism of the prototype concatenated panorama system are also described in this paper.

Teaching

Spring 2006: Visual Geometry Computing

Spring 2005: Visual Geometry Computing

Selected Publications

Other Publications

Professional Activities

Last Modified: August 24, 2007 by Xianghua Ying (xhying@cis.pku.edu.cn )