Lion Based Butterfly Optimization with Improved YOLO-v4 for Heart Disease Prediction Using IoMT

Abstract

The Internet of Medical Things (IoMT) has subsequently been used in healthcare services to gather sensor data for the prediction and diagnosis of cardiac disease. Recently image processing techniques require a clear focused solution to predict diseases. The primary goal of the proposed method is to use health information and medical pictures for classifying the data and forecasting cardiac disease. It consists of two phases for categorizing the data and prediction. If the previous phase's results are practical heart problems, then there is no need for phase 2 to predict. The first phase categorized data collected from healthcare sensors attached to the patient's body. The second stage evaluated the echocardiography images for the prediction of heart disease. A Hybrid Lion-based Butterfly Optimization Algorithm (L-BOA) is used for classifying the sensor data. In the existing method, Hybrid Faster R-CNN with SE-Rest-Net-101 is used for classification. Faster R-CNN uses areas to locate the item in the picture. The proposed method uses Improved YOLO-v4. It increases the semantic knowledge of little things. An Improved YOLO-v4 with CSPDarkNet53 is used for feature extraction and classifying the echo-cardiogram pictures. Both categorization approaches were used, and the results were integrated and confirmed in the ability to forecast heart disease. The LBO-YOLO-v4 process detected regular sensor data with 97.25% accuracy and irregular sensor data with 98.87% accuracy. The proposed improved YOLO-v4 with the CSPDarkNet53 method gives better classification among echo-cardiogram pictures.