A Design Approach for Performance Analysis of Infants Abnormality Using K Means Clustering

The common challenge observed in the early stages of pregnancy is the birth defect of infants. The key factors for this challenge are genetics and infection during pregnancy. According to GHO information, in 2015 about 4.5 million deaths occurred due to the sudden death syndrome and lack of nourishment of the fetus during pregnancy. One of the most important causes for abnormalities in infants is the bulge in their legs and abdomen. Bulge leads to many other problems and affect body functions such as brain, hand, and legs mostly in abdomen. In this paper, 117 images obtained from Beth Israel Deaconess Medical are taken for research purpose i.e., to identify the abnormalities in the fetal brain by using unsupervised learning algorithm. Proposed system is equipped to detect or classify the abnormalities of the fetus having gestational age from 14-38 weeks. Head region and abdomen region of the fetus is used for futher research analysis. Convex hull method is applied to the acquired images for performing image segmentation. The parameters like head diameter and abdomen circumference are used to incorporate feature extraction and followed by that k-means clustering algorithm is used to classify abnormalities in infants. The proposed system gives promising results for detecting the abnormalitiesof fetus and the accuracy is coming out to be 83.76% by using K-means clustering algorithm.