An automated 3D convolutional neural network approach for early alzheimer’s disease diagnosis using brain MRI Scans / Nadjat Manal Boumaza

Public ISBD Titre : An automated 3D convolutional neural network approach for early alzheimer’s disease diagnosis using brain MRI Scans Type de document : document multimédia Auteurs : Nadjat Manal Boumaza, Auteur ; Nardjes Hamini, Directeur de thèse Editeur : Laghouat : Université Amar Telidji - Département d'informatique Année de publication : 2026 Importance : 50 p. Accompagnement : 1 disque optique numérique (CD-ROM) Note générale : Option : Data science and artificial intelligence Langues : Anglais (eng) Mots-clés : Alzheimer’s disease  Deep learning  3D Convolutional neural network (3D CNN)  Magnetic resonance imaging (MRI)  Early diagnosis  Medical image classification  Explainable artificial Intelligence (XAI)  Grad-CAM  Computer-Aided Diagnosis (CAD) Résumé : Alzheimer’s disease (AD) is a progressive neurodegenerative disorder that affects memory, cognition, and daily functioning, making early diagnosis essential for effective clinical intervention and patient management. Recent advances in artificial intelligence have demonstrated the potential of deep learning techniques for the automated analysis of medical images. In this study, an automated framework based on a Three-Dimensional Convolutional Neural Network (3D CNN) is proposed for the early detection and classification of Alzheimer’s disease using brain Magnetic Resonance Imaging (MRI) scans. Unlike conventional approaches that analyze MRI slices independently, the proposed method reconstructs complete MRI volumes by grouping all slices belonging to the same patient, thereby preserving volumetric anatomical information. The reconstructed volumes are processed by a 3D CNN architecture designed to learn discriminative spatial features associated with different stages of Alzheimer’s disease. To improve model transparency and clinical interpretability, a Gradient-weighted Class Activation Mapping (Grad-CAM) module is integrated into the framework, enabling visualization of the brain regions that contribute most significantly to the classification decisions. Experimental evaluation demonstrates the effectiveness of the proposed approach in distinguishing between Non-Demented, Very Mild Demented, Mild Demented, and Moderate Demented subjects. The obtained results indicate that volumetric analysis using 3D CNNs provides a more comprehensive representation of brain structures than traditional 2D methods, while the Grad-CAM module enhances trustworthiness by providing visual explanations of the model’s predictions. The proposed framework represents a promising solution for computer-aided diagnosis systems and may contribute to earlier and more reliable detection of Alzheimer’s disease. note de thèses : Mémoire de master en informatique An automated 3D convolutional neural network approach for early alzheimer’s disease diagnosis using brain MRI Scans [document multimédia] / Nadjat Manal Boumaza, Auteur ; Nardjes Hamini, Directeur de thèse . - Laghouat : Université Amar Telidji - Département d'informatique, 2026 . - 50 p. + 1 disque optique numérique (CD-ROM). Option : Data science and artificial intelligence Langues : Anglais (eng) Mots-clés : Alzheimer’s disease  Deep learning  3D Convolutional neural network (3D CNN)  Magnetic resonance imaging (MRI)  Early diagnosis  Medical image classification  Explainable artificial Intelligence (XAI)  Grad-CAM  Computer-Aided Diagnosis (CAD) Résumé : Alzheimer’s disease (AD) is a progressive neurodegenerative disorder that affects memory, cognition, and daily functioning, making early diagnosis essential for effective clinical intervention and patient management. Recent advances in artificial intelligence have demonstrated the potential of deep learning techniques for the automated analysis of medical images. In this study, an automated framework based on a Three-Dimensional Convolutional Neural Network (3D CNN) is proposed for the early detection and classification of Alzheimer’s disease using brain Magnetic Resonance Imaging (MRI) scans. Unlike conventional approaches that analyze MRI slices independently, the proposed method reconstructs complete MRI volumes by grouping all slices belonging to the same patient, thereby preserving volumetric anatomical information. The reconstructed volumes are processed by a 3D CNN architecture designed to learn discriminative spatial features associated with different stages of Alzheimer’s disease. To improve model transparency and clinical interpretability, a Gradient-weighted Class Activation Mapping (Grad-CAM) module is integrated into the framework, enabling visualization of the brain regions that contribute most significantly to the classification decisions. Experimental evaluation demonstrates the effectiveness of the proposed approach in distinguishing between Non-Demented, Very Mild Demented, Mild Demented, and Moderate Demented subjects. The obtained results indicate that volumetric analysis using 3D CNNs provides a more comprehensive representation of brain structures than traditional 2D methods, while the Grad-CAM module enhances trustworthiness by providing visual explanations of the model’s predictions. The proposed framework represents a promising solution for computer-aided diagnosis systems and may contribute to earlier and more reliable detection of Alzheimer’s disease. note de thèses : Mémoire de master en informatique

---

1 (1 - 1 / 1) Par page : 25 50 100 200