Enhanced Detection of Fetal Congenital Cardiac Abnormalities through Hybrid Deep Learning Using Hunter-Prey Optimization

Congenital heart disease (CHD) is one of the common birth defects, affecting similar to 1% of live births, and is the highest birth defect-related contributor to infant mortality in developing countries. Prenatal diagnoses of critical CHD allow delivery planning for optimal neonatal intervention and medical care, decision-making, and family preparation. Children with prenatal diagnoses are less preoperative brain injury, lower morbidity, more-robust microstructural brain development, and lower mortality for some lesions than those with postnatal diagnoses of CHD. More successful prognoses and better treatment are dependent on earlier detection during the phase of embryonic development. Lately Deep Learning and Machine Learning methods are most commonly used for automatic detection and classification of CHD. This manuscript offers the design of HPOHDL-CHDDF - Hunter Prey Optimization with Hybrid Deep Learning-based Congenital Heart Disease Detection of Fetus (HPOHDL-CHDDF) technique. The goal of the HPOHDL-CHDDF technique is to improve the accuracy and efficiency of CHD detection. To accomplish this, the presented Hunter Prey Optimization with Hybrid Deep Learning-based Congenital Heart Disease Detection of Fetus technique follows two major phases of operations such as Hybrid Deep Learning-based classification and hyperparameter tuning. At the initial stage, the Hunter Prey Optimization with Hybrid Deep Learning-based Congenital Heart Disease Detection of Fetus system involves the design of Convolutional Neural Network with Long Short Term Memory algorithm for classification purposes. Next, in the second stage, the Hunter Prey Optimization with Hybrid Deep Learning-based Congenital Heart Disease Detection of Fetus technique employs the HPO algorithm for optimal selection of the hyperparameter values of the Convolutional Neural Network - Long Short-Term Memory (CNN-LSTM) system. The performance validation of the Hunter Prey Optimization with Hybrid Deep Learning-based Congenital Heart Disease Detection of Fetus algorithm is tested on medical datasets. The experimental values stated that the Hunter Prey Optimization with Hybrid Deep Learning-based Congenital Heart Disease Detection of Fetus technique reaches enhanced performance over other baseline models.