Machine Learning Based Fetal Growth Estimation and Optimization Using Ultrasound Imaging Modality

Abstract

newlineIn modern era, medical imaging has become an important research field in the measurement of diagnosis and treating diseases. The role of present technology is to extract large amount of medical information by utilizing smart and fast techniques that helps in processing of images. Recent advances in medical imaging provide high resolution 3D images that lead to adopt various modern health care strategic principles particularly in the field of medicine. The techniques utilized in medical imaging help doctors in diagnosing the most important public health problems. Estimation of gestational age is one of the medical issue in assessing the high risks existing during pregnancy. Ultra-sonography is an ultrasound newline based diagnostic medical imaging technique that helps a doctor to evaluate, diagnose and treat medical conditions. Ultrasound issued to visualize fetal images to capture its size, structure, and any pathological lesions that provide valuable information for better understanding of the fetal developmental stage. Ultrasound imaging is considered as an important medical diagnostic tool mainly due to its low cost and safety in imaging fetuses. Measurement of biometric parameters such as measurement biparietal diameter of fetal head (BPD) and head circumference play a key role in estimation of gestational age of fetus. However, because ultrasound images are more noisy and due to improper image acquisition measurement of the biometric parameters manually relies on skilled sonographer. The manual parametric determination leads to multiple decisions and causes observational errors. Considering the above mentioned problems in measurement of fetal parameters, the aim of the research is to determine a fully automatic technique for estimation of the gestation age of a fetus accurately by measuring bi parietal diameter of the fetal head. newlineThe proposed technique is based on least square fitting of an ellipse. The head of the fetus is assumed to be elliptical shape. This method utilizes the three step process: