Myocardium extraction in positron emission tomography based on soft computing

Modeling of myocardial contractility using parameterized super-quadric SPECT images

“…We can create a 3D image of the anatomical distribution of the biological processes being studied. We can also use a tracer assay method with a radio-labeled biologically active compound and a mathematical model to describe the kinetics of the tracer as it participates in a biological process [10].…”

“…They applied software to an analysis of myocardial perfusion SPECT as well as brain SPECT and PET data. Behloul et al [10] designed a self-organized radial basis function network to isolate the myocardium from the noisy background of PET images. They used fuzzy sets and fuzziness measures to compute network errors, and then created a visualization of a 3D volume-rendered myocardium.…”

Directional analysis of cardiac left ventricular motion from PET images.

“…PET has enhanced the understanding of the biochemistry of both normal and abnormal functions within the body, providing the means for imaging the rates of biological processes in vivo (BEHLOUL et al, 2001). A great advantage of using PET compared to SPECT is due to the much larger number of elements with low atomic number that undergo positron emission, such as isotopes of Oxygen, Nitrogen, and Fluorine.…”

“…Although segmentation of cardiac PET images may be performed by visual evaluation on a slice by slice basis, this operation is time consuming and suffers from observer bias (BEHLOUL et al, 2001). Automatic, unsupervised image segmentation has an important role to play in medical imaging, but is a challenging task in PET images due to unpredictable object shapes and inconsistent image quality resulting from noise and sampling artifacts (DEDIC; ALLILI; LECOMTE, 2011).…”

“…GUTIERREZ et al (2003) describe an approach to select the LV cardiac MVF from other structures using a global threshold applied to the MVF; however the approach assumes that LV structures have the highest values of velocity and does not account for artefacts such as the spreading of the MVF over image borders due to regularisation, which is a characteristic of the motion estimation method (HORN;SCHUNCK, 1981). BEHLOUL et al (2001) present an approach for myocardium extraction in PET images based on the application of soft computing techniques, while DEDIC; ALLILI; LECOMTE (2011) develop a segmentation method based on a deformable models approach with level sets (LS), where two moving, coupled contours are initialised from inside the left ventricle and outside the heart and deformed to track heart boundaries. The level set (LS) method was initially developed by OSHER; SETHIAN (1988).…”