Inflammation Segmentation Module using Convolution Neural Network for Alzheimer's Disease Detection using Magnetic Resonance Imaging Inputs

Magnetic Resonance Imaging (MRI) inputs have been most commonly used in diagnosing Alzheimer's Disease (AD) in recent years. The patterns of brain functions are deliberately observed in the MRI inputs for efficient AD detection. This article introduces an Inflammation Segmentation Module (ISM) using a convolution neural network (CNN) to augment better AD detection. This segmentation module identifies the textural differences in inflammation regions through training input cross-correlation. The cross-correlation process is reliable in identifying similar and differential textural features to improve the detection accuracy. The CNN is trained using an external dataset and correlated outputs for high precision. This proposed module is reliable in detecting inflammation textures with high precision and less analysis time.