Magnetic resonance brain images individual recognition with PCA

Abstract

Every individual brain is identified as unique by proper consideration of the background for individual difference in the brain functions of the brain morphology. The proposed method is implemented by using structural magnetic resonance imaging brain recognition is performed using segmentation with the Voxel-Based Morphometric (VBM) approach and Feature Extraction(FE) using Principal Component Analysis(PCA). Brain recognition is identified by computing the Euclidean distance among the image pairs, projected into the same subspace. The petite difference in the Euclidean distances is observed between the same subject when scanned twice and it is due to distinct combination of scanners used between test-training image pairs with/without scanner up-gradation. The obtained results of rank identification and receiver operating characteristic curves show that the brain morphology identifies a particular individual with less false acceptance rate.