Automatic classification of neonatal lung disorder in X-ray images

Neonatal bone

Newborn respiratory distress syndrome (NRDS) typically occurs when the neonatal lungs have not produced enough surfactant which is a substance that helps keep the lungs inflated and prevents them from collapsing. Alone side a physical examination and a blood test to measure oxygen saturation in the blood, X-ray imaging is used to diagnose this condition. Treatment after birth may vary depending on the severity of the condition. It includes providing specialist care in a neonatal unit. I may also include provide oxygen to the neonate through a tube or even in some case putting a neonate on a ventilator to support or take over their breathing. On the other hand, Transient Tachypnea (TT) of newborn affects some babies during the first hours of life when they have fluid in their lungs. TT typically affects babies born prematurely above 32 weeks of gestation. This typically makes neonates breath in a fast rate, but it usually lasts for a short period of time. The problem usually goes away without treatment in 3 days or less (Goldsmith et al., 2016). X-ray imaging can be used to image the neonatal lungs and assess the severity of the case by checking the amount of fluid in the lungs. However, in many cases for neonates born after 32 weeks of gestation, distinguishing between NRDS and TT is tricky for some clinicians due to the similar appearance of the lungs in these two conditions. The goal of this study is to investigate the development of machine leaning solution which aims to classify normal, NRDS or TT  from neonatal chest X-ray in preterm born babies.