Luis Roberto Pereira de Paula

Single photon emission computed tomography (SPECT) is a nuclear medicine tomographic imaging technique based on the measurement of spatial distribution of a radionuclide. This technique is widely used in cardiology to assess myocardial perfusion problems related to blood flow in coronary arteries. SPECT images provide better separation of regions of the myocardium and facilitate the location and definition of perfusion defects.

One of the major challenges in SPECT studies is the efficient presentation of information, since a single study can generate hundreds of images of slices to be analyzed. To address this issue, polar maps (also known as Bull's Eye display) are used. Polar maps are built from slices of the left ventricle and provide summarized information of exams in a two dimensional image.

This work presents a method for the segmentation of the left ventricle in myocardial SPECT studies and the construction of polar maps. The segmentation of the left ventricle is performed to facilitate the process of automatic generation of polar maps. The method uses the watershed transform, in the context of the Beucher-Meyer paradigm. To display the results, it was developed an application called Medical Image Visualizer (MIV).