In the medical field, the segmentation of organs and structures in the patient’s body is a very important task to assist the study of morphological and pathological changes of organs. Normally, specialists perform manual segmentation, however it is time consuming, error-prone and observer-dependent. Here the experience of the expertise influences the quality of the ultimate results. This work aims to develop tools for automatic liver segmentation using data acquired with two medical imaging modalities: Computed Tomography and Positron Emission Tomography, in order to improve the way of obtaining volumes of object and to help the clinician in the study of the organ.
Liver segmentation methods were developed for each modality separately and also for the combination of the two modalities. To validate the implemented algorithms, specialists delineated some images for each exam. The results of segmentation algorithms were then compared with the expert reference. The outputs obtained are reasonable and a good starting point for further work.