Brain MRI Hippocompus Segmentation for Alzheimer's Disease Based on Linknet Model

Alzheimer's disease is a neurological disorder marked by progressive cognitive decline and memory loss due to brain cell degeneration. This study aims to segment the hippocampus in MRI scans of patients with Alzheimer's disease using a deep learning approach. The study employs a LinkuNet convolutional neural network architecture based on fully convolutationl in the decoder for hippocampus segmentation on MRI data. The model was trained and tested using MRI scans from 100 and 35 patients, respectively, from the Alzheimer's Disease Neuroimaging Initiative dataset. The performance of the automatic segmentation was assessed by comparing it to manual segmentation using similarity metrics. The segmentation yielded optimal results after 10 iterations. The method achieved a Dice Similarity Coefficient (DSC) of 9938, F1_score of 0.9938, and valid loss of 0.0462 for the test set. The findings demonstrate the potential of the proposed method for effective hippocampus segmentation. The findings demonstrate the potential of the LinkNet-based model as a robust tool for clinical applications in Alzheimer's disease research and diagnosis.