How to Cite: Farsana, W. F., & Kowsalya, N. (2022). Fetal ultrasound image segmentation using dilated multi- scale-LinkNet. International Journal of Health Sciences, 6(S1), 5282–5295. https://doi.org/10.53730/ijhs.v6nS1.6047 International Journal of Health Sciences ISSN 2550-6978 E-ISSN 2550-696X © 2022. Corresponding author: W. Fathima Farsana; Email: afsheensyed84@gmail.com Manuscript submitted: 27 Feb 2022, Manuscript revised: 18 March 2022, Accepted for publication: 09 April 2022 5282 Fetal ultrasound image segmentation using dilated multi-scale-LinkNet W. Fathima Farsana Research Scholar, Research Department of Computer Science and Applications, Vivekanandha College of Arts and Sciences for Women, Elayampalayam, Periyar University, Salem. Dr. N. Kowsalya Assistant Professor, Department of Computer Science, Sri Vijay Vidyalaya College of Arts and Science, Nalampalli, Dharmapuri, Tamilnadu, India. Abstract---Ultrasound imaging is routinely conducted for prenatal care in many countries to determine the health of the fetus, the pregnancy's progress, as well as the baby's due date. The intrinsic property of fetal images during different stages of pregnancy creates difficulty in automatic extraction of fetal head from ultrasound image data. The proposed work develops a deep learning model called Dilated Multi-scale-LinkNet for segmenting fetal skulls automatically from two dimensional ultrasound image data. The network is modeled to work with Link-Net since it offers better interpretation in biomedicine applications. Convolutional layers with dilations are added following the encoders. The Dilated convolution is used to expand the size of an image to prevent data loss. Training and evaluating the model is done using the HC18 grand challenge dataset. It contains 2D ultrasound images at different pregnancy stages. The results of experiments performed on an ultrasound images of women in different pregnancy stages. It reveals that we achieved 94.82% Dice score, 1.9 mm ADF, 0.72 DF and 2.02 HD when segmenting the fetal skull. Employing Dilated Multi-Scale-LinkNet improves the accuracy as well as all the evaluation parameters of the segmentation compared with the existing methods. Keywords---Fetal ultrasound image segmentation, Deep learning, Dilated convolution, Encoder-decoder, Link-Net. Introduction Ultrasound is a useful tool to monitor fetus and mother during pregnancy due to its non-invasiveness and less expenses for imaging while