Imagine a future where a system can detect within a second whether a newborn has jaundice, a disease that affects 60 percent of babies and can lead to brain damage and hearing loss in severe cases. Imagine that the system immediately initiates a treatment regimen and notifies the nurse via text message. All three scenarios are now possible thanks to the efforts of researchers from Iraq and Adelaide.
Engineers from the University of South Australia and Middle Technical University have designed imaging software that can accurately diagnose jaundice in the blink of an eye, automatically turn on blue LED lights to ward off jaundice, and send the diagnosis as a text message to paramedics.
Jaundice is a common disease in newborns, especially premature babies, and there is an excess of an orange-yellow pigment called bilirubin in the blood. It usually resolves quickly when a baby's liver grows enough to expel it from the body.
However, in severe cases of jaundice, due to sickle cell anemia, blood disorders and deficiencies of certain enzymes, the disease is often treated with phototherapy, which uses fluorescent blue light to break down bilirubin in the baby's skin.
University of South Australia remote sensing engineer Professor Javaan Chahl said jaundice was particularly prevalent in developing countries, where there was often no equipment or trained medical staff to treat it effectively.
Professor Chahl said: "Using image processing techniques extracted from the data captured by the camera, we can inexpensively and accurately screen newborns for jaundice before blood tests are performed."
"When bilirubin levels reach a certain threshold, a microcontroller triggers blue LED light therapy and sends the details to the phone."
"This can be done in a second, really, it can have a big impact in severe cases and, if not treated quickly, can lead to brain damage and hearing loss."
The researchers systematically tested 20 newborns diagnosed with jaundice in an intensive care unit in Mosul, Iraq. The second set of data collected 16 images of newborns, of which 5 were healthy and the rest were jaundiced. The system was also successfully tested on four additional mannequins with white and brown skin tones, some with and without jaundice pigmentation.
"Previous research using sensors to find non-invasive ways to detect jaundice has failed. Tried methods are unreliable, costly, inefficient and, in some cases, cause infections and allergies where the sensor needs to touch the skin," he said. Professor Chahl said.
"Our system overcomes these hurdles, instantly detecting jaundice based on a novel digital color representation, enabling high diagnostic accuracy at relatively low cost. It can be used in hospitals around the world and without laboratory facilities and well-trained are widely used by medical staff in medical centers."
The research has been published in the engineering journal Designs.
