Automated Lung Nodule Detection in Computed Tomography Scans Using Support Vector Classifiers

International Journal of Computer Science and Technology | Vol. 9, No. 4, April 2007 | pp. 25–32

DOI: 10.46882/2007/IJCST/090432

Research Article

Title: Automated Lung Nodule Detection in Computed Tomography Scans Using Support Vector Classifiers

Names of Authors: David Miller¹, Alessandro Rossi²

Authors’ Affiliations: ¹Department of Computer Science, Carnegie Mellon University, Pittsburgh, Pennsylvania, USA; ²Dipartimento di Informatica, Università di Pisa, Pisa, Italy

Abstract:
Early identifying of pulmonary lesions in computed tomography (CT) imagery is critical for improving clinical outcomes. This study introduces an automated computer-aided diagnosis (CAD) framework that extracts morphological and texture features to train a Support Vector Machine (SVM) classifier. The pre-processing pipeline filters out background noise patterns and performs 3D threshold segmentation to isolate candidate nodule structures. The system was validated against a reference dataset containing 450 thoracic CT scans. The classification engine achieved a sensitivity rating of 94.2% alongside a low false-positive index of 0.15 cases per scan. Feature vector evaluation completed execution within an average window of t = 1.15 seconds per series, demonstrating a fast and accurate clinical support tool to assist radiologists in high-volume screening operations.

Keywords: Medical Imaging, Computer-Aided Diagnosis, Support Vector Machines, Texture Analysis, Computed Tomography

Manuscript Timeline: Received January 12, 2007; Revised February 22, 2007; Accepted March 16, 2007; Published April 25, 2007

Citation: Miller, D., & Rossi, A. (2007). Automated Lung Nodule Detection in Computed Tomography Scans Using Support Vector Classifiers. International Journal of Computer Science and Technology, 9(4), 25–32.

Subscribe to read the full article: https://www.internationalscholarsjournals.org/subscribe-to-read