Segmentation of pulmonary parenchyma in CT lung images based on 2D Otsu optimized by PSO

Abstract

In this paper 2D Otsu algorithm based on particle swarm optimization (PSO) is proposed to segment CT lung images. This method can extract pulmonary parenchyma from multisliced CT images, which is primary step to detect the pulmonary disease such as lung cancer, tumor, and mass cells. In the automated pulmonary disease diagnosis, image segmentation plays an important role and image analysis result are depends mainly on the effectiveness of segmentation. In CT lung image segmentation traditional 2D Otsu thresholding method plays a vital role. But its main drawback is complex computation and computation time is more. This limits the application of this algorithm in real time diagnosis of medical images. In this paper an improved 2D Otsu algorithm based on particle swarm optimization is developed to reduce the complex computation and computation time. The PSO belongs to evolutionary algorithm family it does not need gradient information. Here PSO is used to find the optimal threshold to segment and extract pulmonary parenchyma in less computation time. Experimental results show that the proposed method gives segmentation results similar to 2D Otsu algorithm. The computation time of 2D Otsu is approximately 90secs/slice and for proposed PSO method is less than 1secs/slice. Thus reducing the total time involved in automated CT lung image diagnosis. The effective performance of proposed segmentation method is evaluated and compared with traditional two dimensional Otsu algorithm.