LAB PEOPLE

 

son.bmpSon Vo (M.S. Student)


Office: EA 723C

 

Phone: (215) 204-6770

 

E-mail: son.vo@temple.edu

 

 

 

Son Vo received his B.S. degree in Electrical Engineering from Temple University, in 2008. His favorite topics include Image Processing, Computer Vision, Multimedia and Mathematics. He hopes one day he could apply what he learned to detect gravitational wave. Currently, he is pursuing the M.S. degree at Temple University with his focusing research on Visual Surveillance.

  Research interests and specialties

■    Visual Surveillance

■    Visible and Thermal Image Fusion

Current research topics

Research Topic : Visual Surveillance

*       Overview

Visual surveillance has many interesting applications in computer vision area. By using a number of cameras we would be able to track and control activities in small and large regions. For example, we could track abandoned objects of customers on a train station or monitor our ships on the sea to see whether they are in the right directions. And there are many more examples which show the usefulness of visual surveillance.

*       Objectives

Research and experiment with video-based visual surveillance techniques. The project covers motion detection, motion analysis, real-time target tracking and image processing

*       Current Progress

The following methods had been employed for motion detection:

1-Background subtraction

This technique is simple to implement and provide fast result. Differences between video frames and reference frame (background) are presented as binary images, and then blobs analysis is utilized to analyze motions and objects.

2-Frame differencing

This technique is based on the difference in intensity between the current video frame and previous frame. The amount of difference will interpretted as motion between frame. However, this technique has difficult in detecting small motions.

3-Spatio Temperal Block-based using PCA reduction technique

Since pixel-based methods in video survellance suffers from noise and instability with light change, block-based method is used to overcome those problems and also serve as a reliable tool for real-time processing of high resolution videos. The addition of PCA technique for dimensional reduction presents a compact framework with the ability to analyze large scale data. This method was developped as in (1).

 

*       Result

son_web_html_m644e4bc.gif

Figure 1 Screen shot of motion tracking using PCA. Video at Temple University

(1) L. J. Latecki, R. Miezianko, V. Megalooikonomou, D. Pokrajac: Using Spatiotemporal Blocks to Reduce the Uncertainty in Detecting and Tracking Moving Objects in Video, International Journal of Intelligent Systems Technologies and Applications (IJISTA), 1(3/4), pp. 376-392, 2006

UP

■      Past Projects

1.   Developing phone-based GRE tutorial and lottery games for people with visual impaired using VoiceXML.

  1. Developing prototype for wireless insulin pump delivery system
  2. Modeling micro-controller using Tanner EDA.