Semi-automatic algorithm for 3D volume reconstruction of inner ear structures based on CT-scan images

Inner ear volume rendering and segmentation are very difficult tasks due to the nature, the size and position of the structure in the temporal bone. However, it is highly useful to reconstruct three-dimensional volume of computed tomography images which will help in determining the anatomy and topographic relationship between various important structures in the temporal bone, as well as the understanding of the anatomical structures of the inner ear. It can be also very useful in making easier the diagnosis of inner ear diseases. We propose a novel approach with minor intervention of specialists, and region growing segmentation, allowing 3D volume reconstruction of inner ear structures form Computed Tomography (CT) images. The obtained results achieve high quality of segmentation and visualization in addition to usability for other systems, such as computer aided diagnosis (CAD) systems.