AICIP Projects

From Aicip
Jump to: navigation, search

Current Projects

  • UTKFace - A Large Longitudinal Face Dataset for Face Aging Study
  • CPSPowerGrid - Achieving High-Resolution Situational Awareness in Ultra-Wide-Area Cyber-Physical Systems
  • SCN - Distributed Solutions to Smart Camera Networks
  • PowerGrid - A Resilient Real-Time System for a Secure and Reconfigurable Power Grid
  • Firmware - Firmware Approaches to Smart Algorithms
  • MSP - Mobile Sensor Platform Assembly

Past Projects

Collaborative Processing in Sensor Networks
The objective of this project is to develop a data sharing middleware that is able to handle multiple distributed data sources and dynamically changing items, and to assist in real-time INFOrmation Dissemination (INFOD) across multiple agencies for homeland security purposes. The ultimate goal of the INFOD model is to support the timely delivery of valuable information.

Sponsored by: DHS, SERRI (Southeast Region Research Initiative), Y-12, and Oracle

This proposal presents a career development plan for integrating research and education. The research component of this program focuses on the study of collaborative signal and information processing (CSIP) algorithms as well as the supporting computing models and protocols in sensor networks. The educational component is aimed at enhancing the newly created computer engineering program of the department through innovative curriculum development, active student mentoring, and by incorporating research findings into classroom teaching. The project consolidates research and educational efforts, responds to the unique challenges presented by sensor networks, and contributes significantly to the PI's career development.

Sponsored by: ECS NSF Electrical & Communications Systems (ECS)

MU-FASHION is a research effort aimed at developing an agent-based infrastructure for multi-resolution, hierarchical, and collaborative signal processing in sensor networks. This infrastructure will include the following innovations: (1) an agent-based sensor fusion architecture that is network-centric as opposed to bring platform-centric; (2) a fault-tolerant, multi-resolution signal integration algorithm that provides real-time, progressively accurate results; (3) a graph-theorectic self-organizing algorithm for multi-tiered sensor networks; and (4) a next-generation lightweight real-time operating system. This work is led by Duke University, in partnership with the University of Tennessee at Knoxville, and Louisiana State University.

Sponsored by: DARPA IPTO SensIT Program

The underwater sensor network has presented a lot of challenging problems, including the energy issue, the scalability issue, the reliability issue, and the real-time response issue. The limited underwater acoustic transmission channel and the slow propagation speed make these issues even harder to solve. Although it is difficult to meet all the objectives at the same time, we can provide performance evaluations and simulations that would suggest a near-optimal solution given certain constraints. In this project, we evaluate different computing paradigms to carry out distributed processing. We will also develop an underwater sensor network simulator (USNSim) using IAI's agent-based ad-hoc mobile network simulation technology. Evaluation metrics including cost, detection accuracy, response time, degree of fault tolerane, etc. This work is led by Intelligent Automation, Inc. (IAI), in partnership with the University of Tennessee at Knowville.

Sponsored by: ONR Sensing and Systems Division through SBIR Program

Mobile sensor platforms that carry imaging, distancing, and wireless communication capabilities are assembled by the students. Image understanding algorithms to help the mobile sensor understand the environment. A group of these sensors can coordinate to better react to or reason about the real-world phenomena. Most of the existing sensor network testbeds are fixed. The few mobile testbeds either have limited processing capability that prevents them from running interesting and practical applications like target tracking, sensor deployment, and other mission-oriented tasks; or that they are too costly for group deployment. We will design and implement a mobile sensor network testbed that can provide sufficient processing capability for both 1-D signal processing and 2-D image processing with low low (<$500 each platform).

Sponsored by: UT SARIF Funding

ATR using Multispectral Imaging
University of Tennessee, in partnership with North Carolina State University, proposes multi-spectral and geometrical based automated target recognition - a research effort aimed at developing a "smart" target recognition system which takes advantage of both the multi-spectral and geometrical information available at different ranges from the target.

Sponsored by: US Army Space and Missile Defense Command

This is a research program to determine the best geometry and best processing algorithms for implementation of multispectral focal-plane array cameras, based on a mosaicked technology, for use in missiles and smart ordanance. This work is led by North Carolina State University, in partnership with the University of Tennessee.

Sponsored by: US Army Space and Missile Defense Command

Medical Imaging with Thermal Infrared
Detecting Breast Cancer from Thermal Infrared Images by Asymmetry Analysis
This is a research effort that helps define thermal infrared (TIR) imaging as a diagnostic tool in breast cancer detection, which can be used as a complementary modality to traditional mammography. This research includes the following innovations: (1) it first proposes a completely automated segmentation and detection algorithm to analyze TIR images of breast; (2) it develops an experimental plan to apply the proposed algorithm to both thermogram and mammogram taken from the same patient of cases including FP, FN, TP, and TN to get objective comparison of the sensitivity of these two imaging procedures in breast cancer detection. Testing images will be provided by Elliott Mastology Center at Baton Rough, LA, and Ville Marie Breast Cancer Center in Montreal, Canada.

Sponsored by: DoD CDMRP BCRP

From Tank to Tumor - Seeing Unseen by Slicing Thermal Texture Maps

Presentation at the Workshop on the Applications of IR Imaging and Automatic Target Recognition Image Processing for Early Detection of Breat Cancer

  • PPT version (including animation)
  • PPT 95 version
  • pdf version


Image Processing and Optimization
Optimal Missing Data Estimation for Large Sensor Arrays
Compared with all the other semiconductor imaging sensors, the charge-coupled-device (CCD) provides higher resolution, wider dynamic range, and higher sensitivity, with which conventional photography can not compete. However, CCD sensors also cost more. Some applications like medical imaging require both high resolution and large area, which makes the device even more expensive. How to achieve large area and high resolution, while at the same time remaining cost-effective is the mainly concern of this research.

Sponsored by: Army Research Office

Work done at NCSU.