An Improved Detection of Fetal Heart Disease Using Multilayer Perceptron

Fetal Heart Disease (FHD) is a major health issue and challenge faced by the entire world in modern medicine. About 25% of babies are affected by heart disorders and 4 out of 10 have died in 2022. The desperation of FHD has become a crucial factor in the increasing rate of mortality and can even result in vulnerable consequences if not predicted at the initial stage. Most medical specialists and practising doctors have found that it's hard to anticipate and identify the disease with the existing traditional techniques at an early stage. This is because of insufficient volume of data, which has become the key failure for predicting this fetal illness and has turned out to be the most challenging task for the medical community. Since the generated data from the human body are continuous and huge in amount, the data mining techniques are utilized for an efficient classification process as it needs an accurate execution of the obtained health data. This paper attempts to reduce the risk of FHD through effective feature selection and classification-based prediction system in Ultrasound (US) images with high-performance measures and accuracy. In this implementation, the input will be obtained from the medical dataset and performs pre-processing followed by the propounded feature selection technique that efficiently decides the selection of features. Based on selected features, a novel classification is performed via Artificial Neural Network (ANN) in the detection of FHD at the early phase. Finally, the efficiency of the system is evaluated using MATLAB where the suggested system possesses an accuracy rate of 98.08% in FHD detection as normal or abnormal in an effective way.