Research On Lung Segmentation Algorithm Based On 2D OTSU Optimized By PSO

Lung disease is a kind of infectious diseases, which is hazardous to human health. In recent years, air pollution is serious, the number of patients is growing dramatically. In order to improve the diagnostic accuracy of lung disease and make the patients to be treated as soon as possible, the ability of computer-aided diagnosis of lung medical images is needed to improve. The segmentation of Lung image is one of key techniques in computer-aided diagnosis of lung medical images. Considering traditional lung tissue segmentation algorithm takes a long time to process, the lung segmentation based on 2D OTSU optimized by PSO is studied in this paper, in order to reduce the operation time of lung tissue segmentation algorithm.Firstly, we study the lung image segmentation algorithm based on CT images. On the basis of understanding CT image-forming principle and the characteristics of CT images, the 2D OTSU algorithm is selected to segment lung tissue in the CT image. Considering the real-time of lung tissue image segmentation, the Particle Swarm Optimization (PSO) algorithm is chosen to optimize 2D OTSU algorithm.Secondly, we study the 2D OTSU algorithm optimized by PSO. In accordance with the shortage of the traditional 2D OTSU based on particle swarm optimization, whose calculating quantity is large and standard particle swarm optimization algorithm is easy to fall into local optimum. The grayscale gradient 2D histogram employed in the paper not only reduces the amount of histogram’s calculation, but also narrows the area which the particles to search. The algorithm also explores the improved PSO which based on diversity of particle symmetrical distribution to search optimal threshold.Last but not least, we study the lung segmentation algorithm based on 2D OTSU optimized by PSO. In the process of algorithm, the region filling algorithm is employed to remove background in order to make the threshold segmentation of lung better. The 2D OTSU algorithm optimized by PSO is selected to segment lung tissue, and the morphology operations to remove noise and repair holes which in the target image. By calculating the area of the connected domain, the two larger connected areas are chosen as regional lung tissue, isolating lungs and trachea. The segmentation algorithm in this paper can segment the lung in CT image automatically and rapidly, remove the trachea effectively, isolate the left and right lung tissue, and can repair the lesion area image.