Multipath Convolutional Neural Network for Brain Tumor Detection (CNN)

Abstract

Brain tumor segmentation is a new automated medical image diagnosis application. A robust strategy to brain tumor segmentation and detection is an ongoing research problem, and the performance metrics of present tumor detection systems are little understood. Deep neural networks employing convolution neural networks (CNN) are being investigated in this regard; however, no generic architecture that can be employed as a robust technique for brain tumor diagnosis has been discovered. The authors have suggested a multipath CNN architecture for brain tumor segmentation and identification that outperforms existing approaches. The proposed work has been tested for datasets BRATS2013, BRTAS2015, and BRATS2017 with significant improvement in dice index and timing values by utilizing the capability of multipath CNN architecture, which combines both local and global paths. It was the objective to provide a simple and reliable method to determine segmentation and detection of brain tumor using brain tumor interface techniques (BCI).