Automatic Extraction of the Lung field from volumetric images for Statistical Anatomical Modeling: A technical approach
Hongliang Ren, Max Q.‐H. Meng · 2011
Statistical Anatomical Models (SAM) of a given population are important to study the anatomical variations of this population. In order to develop SAM models for lung study of a population, we need to extract the anatomical structures of the lung field from a population of volumetric CT datasets. Therefore, it is highly desirable to have an accurate and fast segmentation method without human intervention. This paper presents a fully automatic segmentation method, IBAEL, Intensity Based Automatic Extraction of Lung-field. The segmented lung field structures from the volumetric CT (computer tomography) datasets are used for creating 3D mesh models. Then we can employ a mesh-to-mesh registration method for establishing correspondences and constructing a statistical atlas. The proposed method is aiming to get fast and reasonable segmentations based on conventional image-processing filters for the lung study. IBAEL is the preliminary step for statistical atlas construction, and could be further improved by integrating other prior model based segmentation methods.