Mathematician Linking Statistics, 
Computer Images to Detect Disease


By Ken Keatley

     Carey Priebe's mission as a mathematician is neatly summed
up on the door of his Maryland Hall office, where he's placed a
quote from Leopold Kronecker:
     "The wealth of your practical experience with sane and
interesting problems will give to mathematics a new direction and
a new impetus."
     Dr. Priebe, an assistant professor of mathematical sciences
in the School of Engineering, seeks to apply mathematics and
statistics in solving real problems. Nevertheless, most
acquaintances, upon learning that he's a mathematician,
frequently wonder how he passes his days.
     "They ask, 'What can that possibly be? Hasn't that all been
done?'" Dr. Priebe said. "Unless it's multiplication, they can't
imagine what I do."
     Actually, what Dr. Priebe does would likely surprise many
mathematicians, too. An expert in statistics, he is researching
ways that semiparametric statistical analyses can be applied to
the burgeoning field of computer imaging.
     "These techniques have a number of applications. I'm working
to help radiologists find anomalies in mammograms, but the same
philosophy could help find land mines on a beach or find runways
in satellite imagery," explained Dr. Priebe, who joined Hopkins
in July after a nine-year career in Navy research and development
as a mathematician and scientist. 
     To aid him in this pioneering field of statistical research,
Dr. Priebe has recently been awarded an Office of Naval Research
Young Investigator Program grant. Of 409 proposals received, only
33 were chosen this year. And Dr. Priebe was one of only four
researchers to earn grants in the fields of mathematics or
computer science.
     The three-year grant will provide him with $75,000 in
funding annually, which he plans to use to support the work of
graduate students and to set up in his Maryland Hall office a
sophisticated computer system to test his theories and processes.
     Dr. Priebe, who has degrees in mathematics and computer
science as well as a doctorate in computational statistics, is
excited by the multidisciplinary scope of his work. 
     "One of the main reasons I came to Hopkins is that the
administration truly wants interdisciplinary, collaborative
efforts," he said. "People in electrical engineering, biomedical
engineering and the medical school all have interesting problems
I can try to help solve. I hope I can give some value-added to
them."
     Much of Dr. Priebe's current research focuses on his work in
recognizing and analyzing the texture of mammograms, medical
images used to detect breast cancer.
     "Texture is a statistical measure of the roughness of a
surface," he explained. "By keying in on the texture under
certain conditions, one may find anomalies that may mean the
presence of a tumor or other defect. I'm a statistician, not an
image processing engineer, so I'm looking for statistically
significant departures from the normal tissue. Then I'll try to
draw the attention of a radiologist to those areas of the
mammogram."
     Dr. Priebe is also working with the Navy in developing a
statistical image analysis consulting course that will allow
Hopkins students to solve real-world problems.
     "That should be fun. The advantage for the Navy is to have
some bright Hopkins students working on a problem," he added.
"The students will see some current applications of applied
mathematics."
     Certainly problem solving, in addition to basic research, is
Dr. Priebe's forte.
     "Pure mathematics is beautiful, but it's nice to find a
pay-off somewhere down the line, too," he said.