IRIS Research Student

Wei Hao

Ph.D. Student

  Office: 1414 Circle Dr.
401 Science & Engineering Building
The University of Tennessee
Knoxville, TN 37996-2100
Telephone: (865) 974-9737
Fax: (865) 974-5459
E-mail: whao@utk.edu
Personal Web Page:
Current Research Work

3D Reconstruction Based on Two Wide Baseline Stereo Frames

Recovery of the lost dimension information during the course of imaging is one of the central problems in computer vision. Stereo is an attractive technique for depth perception from multiple views. Small baseline stereo have been investigated in lab and real world applications by many researchers, while there exist certain problems for wide baseline applications. 

We will conduct research in 3D reconstruction using stereo clues. The input of our research will be stereo images of remote scenes from two calibrated cameras. We will devise robust stereo matching algorithms based on geometric invariants, which preserves between images captured from quite different angles.

Based on this research and implementations, a stereo system with high  performance will be set up, and the stereo technique will be applied in related projects in the IRIS Lab.

Multi-Object Motion Pattern Classification and Motion-based Recognition  

Apart from image understanding, video understanding has been a research topic attracting intensive attention from the computer vision community. Motion is one of the most important clues we can employ. In fact, 2D motion information can be used readily for scene segmentation and object recognition in video sequences, no need for the 3D reconstruction before hand, which is supported by the evidences shown by psychologists. Motion pattern recognition based on 2D image sequence is the starting point for further research of video understanding.

Automated video surveillance systems are one of the recent research concentrations of the IRIS Lab. It is an excellent application environment of new techniques in the motion analysis. Automated video surveillance system must be able to track objects moving in its field of view, classify these objects, and understand some of their activities. For instance, in the harbor surveillance system, we are interested in classifying such motions as human activities on the barges and the  background movement such as moving leaves and water ripples.

This research is based on the survey of  various mature and potential techniques in this area. We hope  to build a robust motion pattern recognition system and devise some new techniques of how the motion clue can be used other than the techniques presently available. The system will be built based on some specific surveillance scenes of related projects, and the techniques will be generalized and applied in broader application scopes.

 
 
Schedule:
Monday-Friday: 8:30am-5:30pm
Last updated:
Webmaster