Using computer vision to match the faces of children with undiagnosed intellectual disability.

An estimated 9,800 children locally [17,600 nationally] with intellectual disability (ID) remain undiagnosed, untreated and this casues a devastating impact on families and children. There are many different genetic causes for ID, and most of these are rare. Over 1,500 genes within DNA known to be associated with ID and an estimated further 2,000 ID genes are still to be discovered (out of the estimated 30,000 genes in the human genome). DNA testing of known genes costs thousands of dollars per test, per gene sequence, making it expensive and difficult for parents to obtain diagnoses, especially for rare disorders. 70% of all children with ID remain undiagnosed. 30% to 50% of children with ID have associated skull and facial changes which in a recognisable pattern is referred to as a syndrome. Even with advanced form of genetic testing, less than 30% of children are diagnosed due to the limited  number of syndromes that have been identified.   Currently, clinical geneticists trained in dysmorphology compare photographs of undiagnosed patients at medical meetings in the hope of finding another child with the same rare condition, often likened to finding a needle in a haystack. Current computer generated comparisons are for written descriptions only - which is limited in accuracy. Matching children with similar facial features (morphology) may help identify the genetic cause of their intellectual disability.

Researchers 
Project type 
Project Grant
Year of funding 
2018