Automatic segmentation algorithm for brain MRA images
Muder Almiani, Buket D. Barkana · 2012
In this work, we have developed an automatic segmentation algorithm for brain Magnetic Resonance Angiography (MRA) images to extract the vascular structure based on region-growing method. Intensity information is used as a criterion of homogeneity. MRA was introduced into clinical practice about two decades ago and it provides a variety of significant advantages over competitive methods in vascular imaging. The value of MRA is widely accepted for head and brain imaging. Automatic image segmentation is a prominent process that partitions a digital image into disjoint connected sets of pixels, each of which corresponds to structural units, objects of interest or region of interests (ROI) region in image analysis. The proposed algorithm contains two major stages: (a) Image enhancement and (b) Image segmentation. It provides a parameter-free environment to allow no user intervention. Image denoising and vessel enhancement are useful for improving the display and the segmentation. In order to improve the performance of the region-growing method, we applied contrast enhancement by power-law transformation by the gamma correction technique. Conventional low-pass filter is used as a noise reduction method.