Medical Image Analysis of MRI Brain Images by Deep RBF GMDH-type Neural Network Using Principal Component-Regression Analysis
Tadashi Kondo, Junji Ueno, Shoichiro Takao · 2015
The deep radial basis function (RBF) Group Method of Data Handling (GMDH)-type neural network is applied to the medical image analysis of magnetic resonance imaging (MRI) brain images. In this deep RBF GMDH-type neural network algorithm, many hidden layers are automatically generated and organized so as to fit the complexity of the nonlinear systems by using the heuristic self-organization method which is the basic premise of the GMDH algorithm. This heuristic self-organization method is a type of evolutionary computation. In this study, the deep RBF GMDH-type neural network is applied to the medical image analysis of MRI brain image. The brain regions, the white and gray matter regions in the brain, are recognized and extracted accurately using the deep RBF GMDH-type neural network. These recognition results are compared with those obtained using the conventional sigmoid function neural network trained using the back propagation method.