Health project:New computer programmes are being developed that identify facial features in children to aid the diagnosis of genetic conditions.
Speaking at the press launch of the BA festival of science, Prof Peter Hammond, from University College London (UCL) Institute of Child Health, described his new computer software that can interpret facial features as an aid to diagnosis.
"The software compares the faces of undiagnosed children with those with a diagnosed condition that also affects the development of their face, with a 90 per cent success rate," he stated.
Prof Hammond explained that whilst people recognise the faces of those with Down syndrome, there are, in fact, characteristic facial features of more than 700 genetic conditions.
"For example eyes may be set further apart than usual, the nose shorter and the ears set lower down on the head along with many other possible permutations," he explained.
Although differences in the shape of the face can give clues in early diagnosis before more detailed tests, they can be difficult to detect by the untrained eye. Non-invasive 3D photography and software-based surface modelling techniques developed by UCL aim to make this facial recognition easier.
"Extensive collections of 3D face images of children and adults with the same genetic condition had to be gathered, as well as controls for individuals with no known genetic condition. The images were converted to numeric values to represent each face," said Prof Hammond.
The software will save time, money and reduce stress, as it will narrow down conditions with similar facial features and the diagnosis confirmed with molecular testing.
"Delay in diagnosis causes anxiety to parents who need advice on risks to future children. Moreover, delay may defer important medical treatment or behavioural training that could improve the prognosis for affected children," added Prof Hammond.
"The technique is currently being applied to over 30 conditions with an underlying genetic abnormality. The discriminatory capability of the approach has proven highly accurate in identifying the characteristic facial features of a variety of genetic conditions," he said.