Improve the Detection and Segmentation of Lung Nodule X-ray Images

A combination of three lung nodule detectors has been proposed in this paper, also a new combination of lung nodule segmentation algorithm has been proposed. The proposed detector called Combined Lung Nodule Detector and the proposed segmentation is called Combined Distance Threshold (CDT). The dataset images used in this paper to evaluate lung nodule detection and segmentation is Japanese Society of Radiological Technology (JRST) which consist of 100 malignant cases plus 54 benign cases containing one nodule each. It is clear from results that the results of the CLND detector and CDT segmentation are better than the common traditional methods. As show in Fig. 7 (which represent as prove of concept) that the new proposed detector and the lung nodule segmentation algorithm can enhance the performance of Computer Aided Diagnoses system (CAD) which increase the performance of early detection of cancered lung nodule so can be treated early.