Dr. E.C. Uberbacher, Dr. Ying Xu and Dr. Victor Olman, from the Intelligent Systems Section of the Computer Science and Mathematics Division, detail in their paper, "A Segmentation Algorithm for Noisy Images," their development of a computer algorithm to make it easier to extract objects from an image that contains many different objects. The algorithm, a series of instructions to a computer, can draw boundaries around the object to facilitate object recognition even when the image is noisy or of very low quality.
The research was part of Intelligent System Section's visGRAIL project, a system developed to recognize man-made objects in satellite images, and supported by the DOE Office of Nonproliferation and National Security.
The researchers will be presented the award at the International IEEE Symposium on Intelligent Systems in November in Washington, D.C. The paper will be published in proceedings for the symposium and in an international journal.
Uberbacher started with ORNL in 1991 as Informatics Group leader. Before working at ORNL, Uberbacher was an assistant professor with the Oak Ridge Graduate School of Biomedical Sciences at the University of Tennessee.
Xu, a research staff member, joined ORNL as a postdoctoral research associate in 1993. He was employed as a visiting assistant professor at the Colorado School of Mines prior to his work with ORNL.
Olman, a research associate, has been employed by the University of Tennessee Biology Division since 1993 and began working with ORNL's Computer Science and Math Division in 1995.
ORNL, one of DOE's multiprogram research and development facilities, is managed by Lockheed Martin Energy Research Corp.