Lung Cancer Detection System on Thoracic CT Images Based on ROI Processing

Abstract: The goal is to outline and to develop programmed system for detecting lung cancer and to build efficient outcome within an interactive time frame with least false negative rate. Many systems ensue to increase the accuracy and performance rate. Many imaging techniques are now accessible in the field of medical analysis such as computed tomography (CT), radiography, and magnetic resonance imaging (MRI). Medical image segmentation is very delicate part in analyzing the image. A new approach is used to segment the images and to classify focal areas in lung region. Instead of simple thresholding; adaptive threshold method is applied on dicom images and ROI processing is used to segment the cautious region. Experiments are accomplished using real time datasets to scrutinize our method.

Keywords: Computed Tomography, Gray level Co-occurrence Matrix, Lung Cancer, Minimum False Negative Rate, Neural Network Classifier, ROI, Segmentation.