Segmentation of Lungs in Hrct Scan Images Using Particle Swarm Optimization

A novel segmentation algorithm for lungs based on high-resolution computed tomography (HRCT) scan images is developed. This segmentation method is mainly derived from particle swam optimization (PSO) technique to select an appropriate threshold level for pixels-probability density function (P-PDF) that integrates morphological edgedetection technique to refine segmentation. A multi-level thresholding technique was proposed for CT slice segmentation by developing new control fitness function. After that morphological functions are utilized to get enhanced delineation of lungs from HRCT scan images. For computer-aided diagnostics (CADe) of lungs, this algorithm can be used as an initial step to improve the diagnostic performance of radiologists with an increase in sensitivity and decrease in false-positive (FP) rate. The system was tested on 120 HRCT scan images. This automatic thresholding method was compared with the other state-of-the-art techniques based on ground truth obtained from an expert radiologist. The experimental results indicate that the proposed method provides an effective segmentation solution with small errors, independent of CT scanners and independent from patient’s low or high dose.