A hybrid genetic algorithm and back-propagation artificial neural network based simulation system for medical image reconstruction in noise-added magnetic resonance imaging data

This paper presents, the performance of Genetic Algorithm (GA) applied on a Back-Propagation Artificial Neural Network (BP-ANN) initial weights optimization. The application system is developed for reconstruction of two-dimensional standard magnetic resonance imaging (MRI) images. The BP-ANN initializes the weight between the `ideal' images that are reconstructed using filtered back projection (FBP) technique and the corresponding projection data of simulated MRI image. In an earlier work, it has been reported that as the ANN training time is too long. Hence, in the present work, we propose that the weight datasets are to be processed into GA. Back-propagation artificial neural nets are initialized and GA based weight optimizing is carried out on a system. Consequently, the proposed method makes utilize of `noise-free' projections. When the Genetic BP-ANN strategy is applied, the accuracy of `Noise-added' projections based reconstructed image gets improved. The Genetic BP-ANN approach is able to simplify reconstruction tasks and is seen improving efficiently. The performance results are tabulated for different hidden neuron sizes, population sizes, selection functions, cross-over functions, mutation functions and the number of generations.