Computer Vision News - March 2021

20 Alice Self and Qingchao Chen from the University of Oxford are two of the organizers of this month’s A-AFMA Ultrasound Challenge. Alice is a clinical research fellow in the Nuffield Department of Women’s and Reproductive Health. Qingchao is a postdoc engineering research fellow in the Department of Engineering Science. They speak to us about what to expect from the challenge and how it could help to solve a global clinical problem. The A-AFMA Ultrasound Challenge asks participants to combine simple-to-use scanning protocols with medical image analysis to automate a time-consuming task which usually requires an expert operator. There are two sub-challenges. The first task is to automatically detect two anatomies – amniotic fluid and the maternal bladder – in an ultrasound video. The second task is to measure the amount of amniotic fluid, called the maximum vertical pocket (MVP). “We routinely measure amniotic fluid to give us an idea of fetal wellbeing,” A-AFMA Ultrasound Challenge: Automatic amniotic fluid measurement and analysis from ultrasound video Alice explains. “Both low levels of fluid and high levels of fluid, which are defined as oligohydramnios and polyhydramnios , are associated with a number of fetal and maternal complications, such as fetal growth restriction and maternal diabetes .” An estimate of the amount of fluid surrounding a fetus is a useful clinical guide as to how a pregnancy is progressing . A low level of fluid could mean that the baby’s kidneys are not being adequately perfused, meaning it will struggle, and should either be delivered earlier or at least monitored more carefully during labor. Alice Self A-AFMA Ultrasound Challenge