A new three-dimensional automatic modified region growing algorithm for segmentation of the brain vasculature using magnetic resonance angiography (MRA) image database

Revolutionary progress has been witnessed in medical imaging and computerized medical image processing in the last two decades. The development and advances in multidimensional medical imaging modalities such as magnetic resonance imaging (MRI) have provided important radiological tools in disease diagnosis, treatment evaluation, and intervention for significant improvement in health care.;Within the scope of this study, a comprehensive literature review is performed, and a new medical image processing framework is implemented. In this thesis proposing several computerized segmentation algorithms to extract cerebral vessels using a magnetic resonance angiography (MRA) database that permits the non-invasive visualization of blood flow through the effects of moving spins on the magnetic resonance signal. It is produced using the time-of-flight (TOF) method. The framework described here is based on two major stages: (a) Image enhancement and (b) Image segmentation. In order to improve the performance of the image segmentation stage, image enhancement methods are applied first by the gamma correction technique and spatial-mask processing. This stage is considered to be a part of the framework in the detection of gray-level discontinuities in images. Based on modified region growing method and adaptive maximum intensity difference parameters image segmentation, stage is performed by sliding block-by-block operations.;Because of the small size objects of interest for the blood vessels in each 2D-MRA slice and complex surrounding anatomical structures (e.g., fat, bones, or gray and white brain matter), and the shortcomings typical of sampled data, such as sampling artifacts, spatial aliasing, and noise, which often cause the boundaries of structures to be indistinct and disconnected, new techniques for more accurate segmentation of a 3D cerebrovascular system from TOF-MRA data are proposed. My thesis present that due to the MRA data, blood vessels can be accurately separated from the background in each slice using the now method, the 3D model implemented in my framework. It created a neat mechanism for multi-resolution deformable curve, surface, and solid models to flow or grow into objects with complex geometries and topologies, and adapt their shape to recover the object boundaries. Experiments with real data sets confirm high accuracy of the proposed approach.