An Approach for Fetal Weight Estimation Using Machine Learning for Women Safety

Abstract

Women safety have its importance in society. women should be safer in all aspects. so here is a model which helps them during their pregnancy. The awareness of health in pregnancy is less in uneducated women. We are building a user-friendly model which can be used by clinicians to identify the risks in of fetus whose effect is usually to carrying mother which linked to each other biologically and as its user friendly it can also be used by patients. Our model is built on considering the fetal health and mother too so that we can avoid the life-taking risk of women and the fetus. As in pregnancy one of the important aspects is knowing the weight of the fetus in the womb. It holds importance to clinicians in the management of pregnancy and delivery by keeping track of the mother’s health conditions. As per World Health Organization(WHO), the range of Low Birth Weight(LBW) ranges around less than 2500g, the range of High Birth Weight(HBW) ranges around greater than 4000g, the range of Normal Weight is between 2500g to 4000g. As the fetuses and mothers link to each other biologically, they may both undergo some short and long-term health conditions. To quote some of them are high parent mortality rate, macrosomia as when its High Birth Weight, mental illness in child as long term disorder in Low Birth Weight and chronic diseases in life. To avoid these difficulties we are finding our way in the field of Machine learning and Image Processing concepts. In this paper, we are aiming to develop a model using the Convolution Neural Network(CNN) and a Multiclass SVM algorithm to improve the estimation of fetal weight accuracy and classifying the major disease if the fetus is suffering from a disease which in turn helps clinicians identify the risks before delivery. A balanced dataset is analyzed from zenodo.org. Then these undergo the process of training and classification in two different algorithms for weight estimation and disease classification.