Advanced Classification CT Heart Images Using Second Order PDE Filter and Different Feature Extraction Techniques

Abstract

This paper presents a comprehensive approach to improving the analysis of heart CT images through advanced image processing techniques. We employ  the D2Q9 lattice model for efficient  filtration, specifically tailored to solve the Perona-Malik second-order partial differential equation, a nonlinear diffusion model widely utilized in image denoising and edge enhancement. Additionally, we integrate histogram-based local descriptors to extract meaningful features from the processed images, facilitating more accurate characterization of cardiac structures and abnormalities. Furthermore, the KNN Classifiers utilized to optimize the classification of heart CT images based on the extracted features.