Robust Segmentation of Challenging Lungs in CT Using Multi-stage Learning and Level Set Optimization

Abstract

Automatic segmentation of lung tissue in thoracic CT scans is useful for diagnosis and treatment planning of pulmonary diseases. Unlike healthy lung tissue that is easily identifiable in CT scans, diseased lung parenchyma is hard to segment automatically due to its higher attenuation, inhomogeneous appearance, and inconsistent texture. We overcome these challenges through a multi-layer machine learning approach that exploits geometric structures both within and outside the lung (e.g., ribs, spine). In the coarsest layer, a set of stable landmarks on the surface of the lung are detected through a hierarchical detection network (HDN) that is trained on hundreds of annotated CT volumes. These landmarks are used to robustly initialize a coarse statistical model of the lung shape. Subsequently, a region-dependent boundary refinement uses a discriminative appearance classifier to refine the surface, and finally a region-driven level set refinement is used to extract the fine scale detail. Through this approach we demonstrate robustness to a variety of lung pathologies.

Injured Lung Volume Estimation on Ct Data Using Linear Metrics

“…total healthy healthy eps VV V   . Table 1 shows the results of a study of the model (2). Dashes in the table indicate that the model gives an unacceptable error value, and in this case model cannot be used.…”

“…In case of interstitial lung lesions (pneumonia, alveolar infiltration) ( fig.1) or severe mechanical injury the contour of the lung is often unexpressed or even absent. Various methods to segment the lungs with this pathology and calculate their volume exist [1,2,3], but they are com-Information Technology and Nanotechnology (ITNT-2016) 402 putationally expensive, while achieving error of 10%, and require a lot of initial information and manual image processing, such as: the location of the damage, main points that characterize the lung, etc. Anyway, all of these methods of segmentation aimed at diagnosis lungs disease and the determining of its quantitative characteristics.…”

Interactive Segmentation of Lung Tissue and Lung Excursion in Thoracic Dynamic MRI Based on Shape-guided Convolutional Neural Networks

“…To improve robustness and efficiency, some researchers have utilized a supervised machine learning model to construct lung segmentation systems. For instance, Neil et al [21] designed hybrid model, in which a hierarchical detection network (HDN) was used to detect stable landmarks on the surface of the lungs to robustly initialize a level-set model for delineating lung contours. In [22], an artificial neural network (ANN) was employed to segment the lung tissue in matrix pencil decomposition MRI, achieving a mean value of 0.89 and a standard deviation of 0.03 for Dice similarity coefficient (DSC).…”

Fully automated liver segmentation using Sobolev gradient‐based level set evolution

“…Therefore, LSFs are always signed distance functions, which define the closest distance between each pixel and the zero‐level set. Active contours can merge or break during evolution and handle topological variations (i.e., LSM allows changes of surface topology implicitly). VLSMs can be adapted to several image segmentation problems because they allow the incorporation of several prior information (such as intensity values or shape) to get more robust segmentation under energy minimization framework .…”