Level Set Based Surface Capturing in 3D Medical Images

Abstract

Abstract. Brain aneurysm rupture has been reported to be directly related to the size of aneurysms. The current method used to determine aneurysm size is to manually measure the width of the neck and height of the dome on a computer screen. Because aneurysms usually have complicated shapes, using the size of the aneurysm neck and dome may not be accurate and may overlook important geometrical information. In this paper we present a level set based illusory surface algorithm to capture the aneurysms from the vascular tree. Since the aneurysms are described by level set functions, not only the volume but also the curvature of aneurysms can be computed for medical studies. Experiments and comparisons with models used for capturing illusory contours in 2D images are performed. This includes applications to clinical image data demonstrating the procedure of accurately capturing a middle cerebral artery aneurysm.

Adaptive Contextual Energy Parameterization for Automated Image Segmentation

“…Some form of spatially adaptive regularization over a single image appeared in a recent work by Dong et al [13]. For segmenting an aneurysm, they varied the amount of regularization based on the surface curvature of a pre-segmented vessel.…”

Automatic Intracranial Space Segmentation for Computed Tomography Brain Images

“…Let S ics denote the current ICS segmentation estimate, and let ∅ and d denote signed distance functions with respect to the boundaries of S ics and S bone , respectively. Then, we applied the variational technique proposed by [7], which was based on minimizing the following two-term energy functional:…”

“…This algorithm operates in two sequentially executed phases, which are: (1) a conservative ICS segmentation estimate and a bone tissue segmentation using a novel cascade of grayscale intensity thresholding, binary mathematical morphology, and connected component operations and (2) refinement of the initial segmentation by smoothing the segmentation along low-contrast segments of the image, where bone is absent, in order to approximate the assumed smooth geometry of the ICS boundary. For the second phase, we utilized a modified version of the illusory surface method of [7], which employed a level set technique that drove the segmentation boundary towards the inner boundaries of segmented objects while completing missing borders smoothly. This method utilizes a similar curvaturelimiting mechanism to [6]; however, we modify the method to include a provision to prevent the segmentation boundary from leaking into the brainstem.…”

Towards Self-Parameterized Active Contours for Medical Image Segmentation with Emphasis on Abdomen

“…Although their GUI is friendly to non-expert users, parameter settings are still empirically determined. Dong et al [20] present an algorithm to capture brain aneurysms from the vascular tree, by varying the regularization term based on the surface curvature of a pre-segmented vessel. However, the regularization weight does not rely on image content.…”