[en] Computer images are routinely used in diagnostic centers and hospitals. In particular in the field of Nuclear Medicine they help in the diagnosis and planning therapy against cancer. In the case of the planning therapy the quantifying the distribution of dose in patients is very important, because it provides an estimate of the dose in the tumor and healthy tissues, allowing a greater understanding on the response and toxicity caused by this dose. The aim of this study is to analyze both kinds of images: CT and PET/SPECT and their potential utilization for dosimetry calculation. PET or SPECT images were analyzed using a Gamma Camera, brand Medis, model Nuclide-TH/22 through image acquisition of scanned phantoms containing a known activity inside their volume so that a relationship between the number of counts for each voxel in the image and the real activity will be constructed. The heterogeneous organism patient's is specified from the computed tomography (CT) through number of Hounsfield. However, there is not a simple correlation to convert Hounsfield numbers into material tissues, therefore, in this work we developed a software in Java to convert Hounsfield numbers in mass density. Moreover, the software provides a map of tissues and a text file containing the elemental weights to be used by the Monte Carlo transport code MCNP5 to perform dose calculations. (author)