Due to rapid advancement in medical imaging, human anatomy is now observable in finer details bringing new dimensions to diagnosis and treatment. One such area which benefitted from advancement in medical imaging is aorta segmentation. Aorta segmentation is achieved by using anatomical
features (shape and position of aorta) using specialized segmentation algorithms. These segmentation algorithms are broadly classified into two categories. The first type comprises of fast algorithms which exploits spatial and intensity properties of images. The second type are iterative algorithms
which use optimization of a cost function to track aorta boundaries. Fast algorithms offer lower computation cost, whereas iterative algorithms offer better segmentation accuracy. Therefore, there is a tradeoff between segmentation accuracy and computational cost. In this work, a hybrid approach
for aorta segmentation in 3D Computed Tomography (CT) scan images is proposed. The proposed approach produces high segmentation accuracy of intensity based (fast) approaches at reduced computational cost. The proposed technique is evaluated using real world 3D abdominal CT scan images. The
proposed approach can either be used as a fast-standalone segmentation procedure, or as a pre-segmentation procedure for iterative and more accurate approaches.
PERFORMANCE COMPARISON ON MEDICAL IMAGE SEGMENTATION ALGORITHMS
