Brain Image Segmentation Based on the Hybrid of Back Propagation Neural Network and AdaBoost System

The segmentation of brain magnetic resonance (MR) images can provide more detailed anatomical information, which can be of great help for the proper diagnosis of brain diseases. Therefore, the study of medical image segmentation technology is crucial and necessary. Owing to the presence of equipment noise and the complexity of the brain structure, the existing methods have various shortcomings and their performances are not ideal. In this study, we propose a new method based on back propagation (BP) neural networks and the AdaBoost algorithm. The BP neural network that we created has a 1–7-1 structure. We trained the system using a gravitational search algorithm. (In this algorithm, we use segmented images, which were obtained by state-of-the-art methods, as ideal output data.) Based on this, we established and trained 10 groups of back propagation neural networks (BPNNs) by applying 10 groups of different data. Subsequently, we adopted the AdaBoost algorithm to obtain the weight of each BPNN. Finally, we updated the BPNNs by training the gravitational search and AdaBoost algorithms. In this experiment, we used one group of brain magnetic resonance imaging (MRI) datasets. A comparison with four state-of-the-art segmentation methods through subjective observation and objective evaluation indexes reveals that the proposed method achieved better results for brain MR image segmentation.