Related Experiment Videos
An intelligent diagnostic system for the assessment of gestational age based on ultrasonic fetal head measurements
M S Beksaç1, Z Odçikin, A Egemen
1Department of Obstetrics and Gynaecology, Hacettepe University, Ankara, Turkey.
The object of this study is the assessment of gestational age by using fetal biparietal diameter and head circumference based on automated image analysis and artificial neural networks. Standard ultrasonic measurements were made in 143 normal fetuses between 14 and 40 weeks' gestation. Six hundred and thirteen fetal head images were transferred to a microcomputer environment by means of a frame grabber and spatial software. Biparietal diameter and head circumference measurements were made by automated image processing and analysis techniques. In the next stage, these two fetal parameters were used to determine the gestational age by using an unsupervised artificial neural network. Back propagation learning algorithm was trained by 552 fetal head images and the system was tested with the remaining 61 images. It has been demonstrated that 98% of gestational weeks were estimated correctly by our system.
The object of this study is the assessment of gestational age by using fetal biparietal diameter and head circumference based on automated image analysis and artificial neural networks. Standard ultrasonic measurements were made in 143 normal fetuses between 14 and 40 weeks' gestation. Six hundred and thirteen fetal head images were transferred to a microcomputer environment by means of a frame grabber and spatial software. Biparietal diameter and head circumference measurements were made by automated image processing and analysis techniques. In the next stage, these two fetal parameters were used to determine the gestational age by using an unsupervised artificial neural network. Back propagation learning algorithm was trained by 552 fetal head images and the system was tested with the remaining 61 images. It has been demonstrated that 98% of gestational weeks were estimated correctly by our system.