Left ventricle quantification in gated cardiac nuclear medicine images using Bayes classification

Bayes classification using a Poisson model was developed to classify the gated cardiac nuclear medicine images. Based on the pixel classification, the left ventricle size can be calculated. From the multiple slices of images acquired during a cardiac cycle, the left ventricle volume curve can be obtained and would be available for clinical evaluation. Experimental results showed that the ejection fraction obtained from the volume curve is not very sensitive to variation of the prior probability estimation.<>