pictures may record more than milestone events and the little incidents
we love to remember. New technology may help doctors identify rare
genetic conditions by analyzing ordinary digital photos of faces. Facial
recognition software may even be useful in identifying presently
unknown rare disorders with symptoms that baffle families and doctors.
A New Scientist story by Andy Coghlan
explains that while genetic tests exist for common conditions, such as
Down's syndrome, genetic tests for many more unusual conditions aren't
available because the gene variants that cause them haven't been
discovered. A story in The Independent by Charlie Cooper explains that 30 to 40 percent of genetic disorders involve some kind of change to the face or skull. Software developed at Oxford University
by medical researchers collaborating with the university's Department
of Engineering Science was initially "trained" by analyzing thousands of
photos of people diagnosed with eight genetic disorders. Coughlin's
story explains that the computer "learned" to identify each condition
from a pattern of 36 features in each face.
who designed the software with Oxford colleague Andrew Zisserman,
believes it can help family doctors and general pediatricians make
preliminary diagnoses of health conditions that may have puzzled them.
In the future, Nellaker told The Independent, a doctor anywhere
in the world should be able to take an ordinary smartphone picture of a
patient, run a computer analysis, and find out which genetic disorder a
patient is likely to have. The technology isn't meant to replace
traditional diagnoses, but to aid it by giving doctors information not
otherwise available to them.
Alastair Kent, director of the Genetic Alliance UK, a charitable organization dedicated to helping people with genetic disorders, told New Scientist
that because few physicians are skilled in the diagnostic use of facial
analysis, families often wait years to learn the cause of their
children's problems. Many of the combinations of facial characteristics
that have diagnostic significance would be undetectable to a layman.
Oxford database now has nearly 3,000 photos, and the software can
recognize 90 disorders. As the database grows, the software will enable
researchers to study groups of patients with undiagnosed problems who
share similar facial features and skull structures. That could allow
researchers to identify presently unknown disorders and the explore the
gene variants that cause them, which could potentially improvement