Early Alzheimer's Disease Detection Through YOLO-Based Detection of Hippocampus Region in MRI Images

Magnetic Resonance Imaging (MRI) is currently one of the most promising tools for detecting Alzheimer's disease (AD), as it allows for the analysis of brain regions affected by the disease, such as the hippocampus. However, the availability of labeled datasets for hippocampus regions in MRI images is limited, and manually annotating such images can be expensive and time-consuming task, particularly for large datasets. To overcome this challenge, we propose a deep learning approach that leverages object detection models to automatically identify the hippocampus region in MRI images. In our study, we employed various YOLO-based models to detect and classify the AD classes based on the hippocampus region in MRI images. We specifically selected the latest state-of-the-art YOLO variants, including YOLOv3, YOLOv4, YOLOv5, YOLOv6, and YOLOv7. Our approach shows potential for improving the early detection of Alzheimer's disease using deep learning and object detection and may be useful for developing automated diagnostic tools for clinical applications. We conducted experiments in two scenarios to validate our proposed idea: one-class detection and two-class detection. One-class detection detects a specific class based on the appearance of the hippocampus region, while two-class detection aims to detect and classify the AD level based on the hippocampus. Our preliminary results demonstrate that YOLO variants are viable for accurately detecting the hippocampus region in MRI images, with potential applications in hippocampus detection.