Brain Tumor Classification Based on Attention Guided Deep Learning Model

Abstract Cancer is the second leading cause of death worldwide. Brain tumors count for one out every four cancer deaths. Providing an accurate and timely diagnosis can result in treatments. In recent years, rapid development image classification has facilitated computer-aided diagnosis. The convolutional neural network (CNN) most widely used models classifying images. However, its effectiveness limited because it cannot accurately identify focal point lesion. This paper proposes a novel brain tumor model that integrates attention mechanism multipath to solve above issues. An select critical information belonging target region while ignoring irrelevant details. A assigns data multiple channels, before converting each channel merging results all branches. equivalent grouped convolution, which reduces complexity. Experimental evaluations on this using dataset consisting 3064 MR images achieved overall accuracy 98.61%, outperforms previous studies dataset.