Bibliography

(eng): Factor analysis is a well established mathematical method for factor separation in the analysis of scintigraphical sequences. The results are typically an input to the next step, e.g. factor analysis for computing significant diagnostic coefficients. However, this computing highly depends on proper identification of factors and their biological meaning, which is not ensure only by factor analysis. The main issue is separation overlaping factors from themselves and from tissue background covering the whole sequence. Factor analysis highly depends on prior information which allows us to set biologically reasonable conditions to a mathematical model. In this paper, we propose a mathematical model which estimates the probability mask of each image factor and sets it as a prior information for the next step of iterative algorithm based on Variational Bayes method. The new proposed model provides more realistic estimates of factors than the standard factor analysis.