Enhancement of ultrasound image processing techniques for automated fetal anomaly detection

Abstract

Ultrasound imaging is preferred in obstetrics because it is non-invasive more economical and safe The growth of the fetus should be monitored during the entire term of pregnancy Fetal growth monitoring can be done by measuring various anatomical parameters such as Bi-Parietal Diameter BPD Head Circumference HC Abdominal Circumference AC Occipital Frontal Diameter OFD and Femur Length FL A wide range of fetal malformations and anomalies such as defects in central nervous system heart anterior abdominal wall can be diagnosed from the biometric measurements At present radiologists manually delineate the fetal structures in ultrasound images and make measurements which are time consuming and error prone The images are affected by artifacts such as speckle reverberation ghost mirror shadow noises In addition the quality of the images also depends on: the structure to be imaged Body Mass Index BMI of the mother and gestation age Thus automated fetal ultrasound segmentation and measurement system is the need of the hour The objective of the research is to develop enhanced 2D fetal ultrasound image processing techniques for automated anomaly detection The research carried out in this regard is; pre-processing of ultrasound images speckle noise reduction segmentation of fetal images and automated fetal standard plane detection A speckle reduction technique based on the trilateral filter and local statistics of the image has been developed The local speckle content of the image influences the trilateral filtering newline