AUTHOR=Chen Xijie , He Miao , Dan Tingting , Wang Nan , Lin Meifang , Zhang Lihe , Xian Jianbo , Cai Hongmin , Xie Hongning TITLE=Automatic Measurements of Fetal Lateral Ventricles in 2D Ultrasound Images Using Deep Learning JOURNAL=Frontiers in Neurology VOLUME=Volume 11 - 2020 YEAR=2020 URL=https://www.frontiersin.org/journals/neurology/articles/10.3389/fneur.2020.00526 DOI=10.3389/fneur.2020.00526 ISSN=1664-2295 ABSTRACT=The measurement for the width of fetal lateral ventricles (LV) in prenatal ultrasound (US) images is essential to antenatal neuronographic assessment. However, the manual measurement of LV is highly subjective and relied on the clinical experiences of scanners. To deal with this challenge, we propose a computer-aided detection (CAD) framework for automatic measurement of the fetal LV in two-dimensional (2D) US images. Firstly, we train a deep convolutional network on 2400 images of LV to perform pixel-wise segmentation. Then, pixels-per-centimeter (PPC), a vital parameter for quantifying the caliper in the US images, is obtained by morphology operations guided by the prior knowledge. The estimated PPC, by converting into physical length, is used to determine the diameter of LV by employing the minimum enclosing rectangle (MER) method. Extensive experiments on the self-collected dataset demonstrate that the proposed method achieved superior performances by comparing it with the manual measurement. It yields a mean absolute measurement error of 1.8 mm. Our method is fully automatic and is shown to be capable of alleviating the measurement bias caused by improper US scanning.