Liver Segmentation on CT Images. A Fast Computational Method Based on 3D Morphology and a Statistical Filter
Fernando López-Mir, Pablo González, Valery Naranjo, Eugenia Pareja, Mariano Alcañíz, Jaime Solaz · IWBBIO · 2013
The purpose of this paper is the segmentation of the liver tissue in computed tomography (CT) images and the comparison with several literature methods. Several expert radiologist restrictions such as automation, an easy user-interaction, and a low time-cost were taken into account for selecting the nal algorithm. Thirty public dataset have been used to estimate the accuracy of the algorithm, twenty for training and ten for testing our method. A Jaccard index of 0:89, an average distance of 2:06 mm, and a runtime of 0.54 seconds per image state a promising eciency but a poor ecacy. For this reason, user-assisted tools were used in a nal step to demonstrate that this fast computational method with minimal user corrections (in some required dataset) remains runtime ecient and enough accurate for the liver segmentation purpose.