Automatic Parameterization of Grey-Level Hit-or-Miss Operators for Brain Vessel Segmentation

Abstract

Reliable segmentation of 3D magnetic resonance angiography (MRA) is fundamental for planning and performing neurosurgical procedures, but also for detecting vascular pathologies. We propose here a method for brain vessel segmentation based on mathematical morphology tools. This method, devoted to phase-contrast MRA (PC-MRA) performs vessel segmentation by applying an adaptive set of grey-level hit-or-miss operators on each point of the MR data. High level anatomical knowledge modeled by a vascular atlas is used in order to adapt the parameters of these operators (number, size, and orientation) to the current position. The method has been performed on 30 PC-MRA cases composed of both phase and magnitude images. The results have been validated and compared to segmented data obtained by applying a region-growing algorithm on the same database. They tend to prove that the method is reliable for brain vessel detection and additionnally provides information on vessel size and orientation without requiring any post-processing step. PC-MRA IMAGINGPC-MRA is a non invasive 3D MR technique [9] enabling to generate two different images during a same acquisition. The first data, named magnitude image, contains anatomi-

Grey-level hit-or-miss transforms—Part I: Unified theory

“…In Part II of this paper [29] we will present some applications of the grey-level HMT in the specific case of analysing 3D angiographic image (i.e. medical images visualizing vessels) [25][26][27][28]. This should convince the reader of its wide applicability in the field of grey-level image processing.…”

“…Note also that the special interpretation of the case (F⊖V)(p) = (F⊕W * )(p) = +∞ in (24), and of the case (27), which arose because ±∞ T ′ , does not apply here for (F ⊖V)(p) = (F ⊕ W * )(p) = t 1 or t 0 . Finally, in the binary mask valuation M, the resulting values +∞ and −∞ should be replaced by t 1 and t 0 .…”

“…The authors have successfully applied grey-level HMTs to the detection of blood vessels in 3D angiographic images [25][26][27][28]. In fact, we used both Ronse's and Soille's unconstrained versions, but also some new variants.…”

Watershed and multimodal data for brain vessel segmentation: Application to the superior sagittal sinus

“…More sophisticated mathematical morphology operators have recently been proposed, as in [15,16], where watershed segmentation is performed on cerebral MRA before classifying the obtained primitives by using fuzzy logic techniques or neural networks. Finally, grey-level hit-or-miss operators have been considered too for vessel segmentation applications, in [17,18] for cerebral MRA analysis and in [19] for liver vasculature detection from CT-scans.…”

“…Except in very few examples such as [14] and [17,18] where knowledge about vascular structure topology and about size and orientation of the searched vessels are used to guide the segmentation processes, nearly all the described methods are based on quite simple hypotheses generally linked to the blood signal intensity and the cylindrical shape of the vessels.…”