Caps Net Based Deep Learning Model with CHO Optimization for Cardiovascular Disease Diagnosis

Cardiovascular disease (CVD) is the most general cause ofsick and mortality in the world. CVD disorders include heart attack, stroke, heart failure, hypertension, atherosclerosis, and ischemic heart disease. The hazardissues of this disease including obesity, diabetes, high blood pressure need to be identified in early for their effective controlling and there are many ways to detect CVD diagnosis. This paper presents CapsNet Transformer based deep learning method for CVD diagnosis. The process begins with data acquisition followed by robust preprocessing is carried out by image resizing and normalization and the segmentation is performed using OTSU threshold method to isolate regions of interest. Subsequently feature extraction process extracts texture, shape and intensity features from medical images using Histogram Oriented Gradients (HOG) technique. The most relevant input featureselection is employedCat Hunting Optimization (CHO) optimization to identify the most discriminative features. These feature selection fed to a classifier of Capsule Networks (CapsNet) and Transformermodel for classification in CVD for detecting heart abnormalities. Finally the predicted output and its performance is evaluated to ensure high efficacy in classification. Consequently the experimental results are verified and obtained from Python software and the overall accuracy achieved is 𝟗𝟔% respectively.