North Carolina A & T State University

Current Research Projects in NC A&T

Numerous research projects are conducted using the IMIS and CRI infrastructure (NSF, DARPA, NSA, TI, VoiceMethod). Some of the research areas are as follows:

• Wireless Networks

- Wireless Sensor Network
- Wide-band CDMA


• Panoramic Image/Video streaming over network

- Region Scalable coding
- Panoramic Image Based Rendering
- Content Tracking and Extraction


• Sensor Network Technology

- Motion tracking (Multiple objects)
- Network Security


• Image/Video Compression



• Other researches

- Automatic Speech recognition and evaluation
- Intelligent Multi-agent Systems

Wireless LAN Infrastructure

We installed wireless LAN extensions for users in the IMIS Lab and IPCV Lab for indoor research. We also installed wireless LAN extensions in the two parking lots between McNair Hall, Graham Building and IRC Building for outdoor research. This cell coverage is to support wireless communication research projects.

Outdoor Access Points
IPCV Lab		IMIS Lab

Mobile Robot with Wireless Communication

The objective of this activity is to support teleoperation research. The mobile robot is equipped with wireless LAN, stereo camera, and sensors. We tested our region scalable video coding algorithm over wireless communication links between the IRC Lab and the mobile robot. We were able to encode and send stereo video from the mobile robot to the IRC Lab. We were able to control the mobile robot based upon the video inputs. This mobile station will be used for various research projects to support the goals of IMIS.

Networked Video Surveillance System

The objective of this research project is to develop a networked surveillance system. We developed a real-time software codec and installed it in one of the computer labs in the Computer Science Department. The functionality of the software included video streaming, device control, and video storage.

Sensor Network Research (I)

The objective of this research project is to develop Sensor network devices for applications such as security, inventory control, etc.

Sensors for Automatic Speech Recognition

The objective of this research is to address the testing and analysis of various microphones versus the Physiological Microphone in different techniques under different environments with the use of a selected ASR program. We are exploring the use of different techniques and environments in which a user interfaces a selected ASR program. The testing of multiple microphones has provided us with varied results based on the environment. The software of choice for our research is Dragon Naturally Speaking 5.0.

P-mic		System Config.		Test results

Wireless Sensor Network for Human Behavior Analysis

Goal is to use innovative sensor network technology to automate assessment of behavior of trainees in a large-scale training exercise.

Patrol Street		Firing Posture

Human Motion Analysis with Panoramic video

Security Algorithm for Sensor Networks

The proposed encryption system seems a very good balance between performance and security given the limits of network sensors.

- RC4 based solution, yields better performance on Atmega128, than most other block ciphers.

- Provides adequate security for most applications.

Neural Network Based Face Verification

Research Objective: Development of robust face tracking/detection techniques for advanced ASR system

Video communication in Heterogeneous Networks

The objective of this research project is to develop a coding algorithm and a corresponding high performance, special purpose computer system to implement video compression and transmission over heterogeneous or wireless communication channels.

Region Scalable Image Compression

Panoramic Image Coding for Virtual Reality Applications

Panoramic Imaging Rendering and Tracking

The objective of this research project is to develop realistic panoramic image rendering techniques. This includes developing a technique for segmenting moving objects in a image scene.? A panoramic image compression technique will also be developed.

Results

(a)		(b)
		(a) original image (b) reconstructed image with basic quality (c) reconstructed image with object region updated
(c)