Pulmonary Artery Blockage Detection

Assistant Professor Senthil Kumar, Aswini. P, Durshika.K.J, Vinusha.V.S

Abstract- – This study investigates the application of Machine Learning (ML) techniques for automated detection of pulmonary artery blockage using medical imaging data. Pulmonary arterial obstruction represents a potentially life-threatening condition requiring rapid and accurate diagnosis for optimal clinical outcomes. We present a novel computational approach that leverages advanced ML algorithms to analyze and interpret pulmonary angiography images. Our methodology incorporates comprehensive image preprocessing, segmentation, and feature extraction techniques to prepare data for classification. Multiple ML architectures were implemented and comparatively evaluated, including Convolutional Neural Networks (CNNs) and Sup- port Vector Machines (SVMs). Performance assessment utilized standard metrics including accuracy, sensitivity, specificity, and area under the receiver operating characteristic curve. Results demonstrate that our ML-based approach achieves superior diagnostic accuracy compared to conventional methods, with potential applications as a clinical decision support tool. This research contributes to the evolving landscape of computer-aided diagnosis in pulmonary vascular pathology and offers promising avenues for improving early detection and treatment planning for patients with suspected pulmonary artery blockage.

DOI: /10.61463/ijset.vol.12.issue2.695