TY - GEN
T1 - Automatic measurement of the fetal abdominal section on a portable ultrasound machine for use in low and middle income countries
AU - Khan, Naiad Hossain
AU - Tegnander, Eva
AU - Dreier, Johan Morten
AU - Eik-Nes, Sturla
AU - Torp, Hans
AU - Kiss, Gabriel
N1 - Publisher Copyright:
© 2016 IEEE.
PY - 2016/11/1
Y1 - 2016/11/1
N2 - General pregnancy monitoring with ultrasound includes the measurement of the biparietal diameter (BPD), femur (FL) and the mean abdominal diameter (MAD) of the fetus. This study was aimed to develop an automatic method for localization of the presented section through the abdomen and measurement of MAD. The algorithm may be run on a conventional ultrasound machine or on a tablet-based machine designed to be used in the rural areas of low- and middle-income countries (LMIC). The open source computer vision (OpenCV) library was used to process a B-mode ultrasound image to detect the location of a fetal abdomen. A Kalman based tracker (RCTL, GE Vingmed Ultrasound) was used to converge a deformable circle model along the boundary of the detected abdomen. The mean of the model's major and minor axes was considered as MAD. The method automatically localized the abdomen and measured MAD in 57 of 61 images (93%) collected from 16 - 41-week-old fetuses. The correlation and the error plots between reference and automatic MAD measurements are presented. The correlation coefficient was 0.96, mean error was -0.06 mm and the error range (95% CI) was -14.80 to 14.68 mm. The resulting errors increased with the gestational age (GA) of the fetuses due to inner shadow and lack of edge information. The method was tested to perform well both in a PC and on various mobile devices.
AB - General pregnancy monitoring with ultrasound includes the measurement of the biparietal diameter (BPD), femur (FL) and the mean abdominal diameter (MAD) of the fetus. This study was aimed to develop an automatic method for localization of the presented section through the abdomen and measurement of MAD. The algorithm may be run on a conventional ultrasound machine or on a tablet-based machine designed to be used in the rural areas of low- and middle-income countries (LMIC). The open source computer vision (OpenCV) library was used to process a B-mode ultrasound image to detect the location of a fetal abdomen. A Kalman based tracker (RCTL, GE Vingmed Ultrasound) was used to converge a deformable circle model along the boundary of the detected abdomen. The mean of the model's major and minor axes was considered as MAD. The method automatically localized the abdomen and measured MAD in 57 of 61 images (93%) collected from 16 - 41-week-old fetuses. The correlation and the error plots between reference and automatic MAD measurements are presented. The correlation coefficient was 0.96, mean error was -0.06 mm and the error range (95% CI) was -14.80 to 14.68 mm. The resulting errors increased with the gestational age (GA) of the fetuses due to inner shadow and lack of edge information. The method was tested to perform well both in a PC and on various mobile devices.
UR - http://www.scopus.com/inward/record.url?scp=84996482893&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84996482893&partnerID=8YFLogxK
U2 - 10.1109/ULTSYM.2016.7728557
DO - 10.1109/ULTSYM.2016.7728557
M3 - Conference contribution
AN - SCOPUS:84996482893
T3 - IEEE International Ultrasonics Symposium, IUS
BT - 2016 IEEE International Ultrasonics Symposium, IUS 2016
PB - IEEE Computer Society
T2 - 2016 IEEE International Ultrasonics Symposium, IUS 2016
Y2 - 18 September 2016 through 21 September 2016
ER -