Segmentation of Lung Region for Chest X-Ray Images Based on Medical Registration And ASM
Chunyan Wang, Shengwen Guo, Xiao-Ming Wu · 2009
It's very important to locate and recognize the lung region accurately in chest X-ray images in clinical application and research. This paper provides a novel method to extract the lung region in the chest X-ray images. In this paper, the active shape model (ASM) based on deformed technology is applied to segment the lung region. In order to get a more accurate and time-saving segmentation result, we also improve the original ASM. Firstly, Because of the body location and the individual difference, we use thin-plate spline method to register the chest radiographs in order to get a more appropriate shape model .Secondly, apex pulmonis and gulus costa are localized and used to initialize the mean shape mode, and the mean intensity of the bound rectangle regions of the initial shapes is adopted to match the shapes. Finally, a multi-resolution framework based on Gaussian pyramid was introduced to acquire quick iteration. Experimental results show that our algorithm performs significantly better than the original ASM.