Search algorithm of a maximum rectangular region for object description

Monte Carlo methods are becoming important in evaluating data correction methods in nuclear medicine. In the Monte Carlo methods, time-consuming calculations are required to determine the path length of a photon in an object that is composed of several media with complicated boundaries. Object modeling is therefore important for performing photon transport efficiently. The authors have proposed a new algorithm called the maximum rectangular region (MRR) method. The results of simulation showed that the calculation time in MRR was about from 30% to 50% of that in the conventional algorithm. But the disadvantage with this algorithm is that it takes too long a calculation time to decide MRR for each voxel. In this paper the authors propose a new method by using a genetic algorithm to decide the MRR for each voxel. The results showed that the calculation time could be reduced by 30% of the conventional algorithm.