This dissertation presents the development and results of this M.Sc project, whose multidisciplinary objective, was to develop a methodology and a tool to segment the liver, its vessels and subregions from abdominal computed tomography images, using procedures of automatic image segmentation and visualization of three-dimensional data. The suggested methodology segments initially the liver, using an approach based on implicit deformable models, called level sets, estimating its parameters using genetic algorithms. Initially, the liver boundary is manually set in one slice an initial solution, and then the method automatically segments the liver in all other slices, sequentially. Then the vessels and nodules of the liver are identified using both a model of mixture of functions proportional to Gaussians, and a segmentation method called region growing that uses hysteresis information. The hepatic and portal veins are classified within the set of vessels, and used in a mathematical modeling that eventually divides the liver into the eight subregions of Couinaud. The methodology was tested to segment the liver using 20 different exams and five different measures of performance, and the results obtained confirm the potential of the method. The cases in which the method presented a poor performance are also discussed in order to instigate further research.