Computer Vision News - March 2021

22 A-AFMA Ultrasound Challenge “The anatomy’s appearance is different and there are fewer useful frames,” Qingchao explains. “We want to see how our challenge participants design their algorithms to analyze this simple-to-use protocol view. Also, the usual MVP measurement process is very time consuming, so we would like to see if computer-assisted algorithms can help this procedure to estimate better and faster.” Something else researchers need to be aware of is avoiding taking measurements of fluid that is not amniotic fluid, such as through any fetal parts or the fetal cord, which is an easy mistake that algorithms could make. The team hope results will mimic those of the most experienced sonographers, which will be the ground truth for the challenge. Alice and Qingchao are keen to give a shout out to their co-organizers, including Professor Alison Noble and Professor Aris Papageorghio . Outside of the challenge, they are all working together on the CALOPUS project, which is an international multidisciplinary project involving a team from the Translational Health Science and Technology Institute (THSTI) based in Delhi, India. “It could be performed by a traditional birth attendant, or a community midwife who wouldn’t normally have ultrasound training.”