Enhancing diagnostic accuracy of fetal chromosomal abnormalities using automated ultrasound image processing of the nasal bone

Abstract

This paper presents an algorithm for the automatic processing of ultrasound images of the fetal head, aimed at detecting and measuring the nasal bone. Ultrasound examination during the first trimester (11–14 weeks) plays a critical role in assessing the risk of chromosomal abnormalities. However, visualizing the nasal bone can be challenging due to its small size and variability in image quality across ultrasound devices. The proposed algorithm combines Gaussian and median filtering for noise suppression and employs an improved Otsu segmentation method to enhance reliability. The system was validated using phantom-generated images as references, achieving a measurement accuracy within one pixel.