Alzheimer's Disease Detection Using Bidirectional Empirical Mode Decomposition

• Published in: International Conference on Signal Processing and Communication (Online) pp. 625 - 630
• Main Authors: Garg, Neha, Garg, Amit, Choudhry, Mahipal Singh
• Format: Conference Proceeding
• Language: English
• Published: IEEE 20.02.2025
• Subjects: Accuracy; Alzheimer's disease; Alzheimer's disease detection; Bidirectional Empirical Mode Decomposition; Empirical mode decomposition; Feature extraction; Feature Selection; Gray Level Co-occurrence Matrix; Magnetic resonance imaging; Matrix decomposition; Medical treatment; Particle swarm optimization; Sensitivity; Signal processing
• ISSN: 2643-444X
• DOI: 10.1109/ICSC64553.2025.10967815

Alzheimer's disease is a progressively deteriorated, irreversible brain ailment that impairs memory and cognitive function. Early detection of Alzheimer's disease is crucial as it can reduce the risk factors and can provide the effective therapy to the affected person within stipulated time. In the proposed approach, bidirectional empirical mode decomposition is applied on 2-D Magnetic Resonance Images to decompose them into four Intrinsic Mode Functions. Then, using four Intrinsic Mode Functions, the Gray Level Co-occurrence Matrix has been computed. The Gray Level Co-occurrence Matrix interacts directly with the image intensity, it offers information on patterns that mutually occur. For every Intrinsic Mode Functions, 22 statistical features are obtained from these matrices. In order to improve the classification accuracy, Particle Swarm Optimization (PSO) is further utilized for extracting optimal and relevant features. The classification results for proposed method are 96.4% accuracy, 96.0 % sensitivity and 97.0 % specificity. The proposed work outperforms the existing methods for AD detection in terms of its outstanding classification results.