Convolutional Neural Network to Detect and Measure Fetal Skull Circumference in Ultrasound Imaging

Abstract

In obstetrics, ultrasound is used for assessment of fetal development during pregnancy. The images generated by ultrasound are used to obtain measurements of fetal head length, body size, and the analysis of fetal movements, to identify and prevent the onset of congenital disease. This work presents the development of a new method for the segmentation of two-dimensional ultrasound images of fetal skulls based on a V-Net architecture called Fully Convolutional Neural Network - Combination (VNet-c). We created a new combination of strategies using a 3D V-Net as base, such as pre-processing, use of Batch Normalization and Dropout, and evaluation of distinct activation layers, activation function, data augmentation, loss function, and network depth. The computational results reveal the feasibility of the proposal in the correct segmentation of fetal skulls and head circumference measurements, reaching up to 97.91% of correctness, overcoming states-of-the-art methods.

Keywords

• Convolutional neural networks
• deep learning
• fetal skull
• measurement