Processing of MR Slices of Human Liver for Volumetry

摘要

This article deals with a segmentation of MR images of human liver. The aim of the segmentation is bounding of tumors in MR images of the human liver in each slice. It is possible to reconstruct the original shape of the tumor from these bounded areas and if we know the thickness of each slice the volume can be solved. This method is suitable for monitoring the development of tumor volume during the treatment. The volume reducing means successful treatment. There was chosen an active contour method based on level set partial differential equation for the MR image segmentation [1]. There was found the boundary of the liver tumor in the noised image of low contrast with suitable parameters and initial curve without any kind of preprocessing. The article shows our results of the segmentation of some MR slices with visible liver tumors. These results are poised to following postprocessing which will be the 3D model creation and the volumetry of the tumor.