| Titre : | AI-Powered PlatformFor E-Health and Paludism Diagnosis Using Deep Learning |
| Auteurs : | Baissi Fadia, Auteur ; Abdelkader Abdelbaki Elhadj, Auteur ; Amira Mohammedi, Directeur de thèse ; Laïd Kahloul, Directeur de thèse ; Asma Ammari, Directeur de thèse |
| Editeur : | Biskra [Algérie] : Faculté des Sciences Exactes et des Sciences de la Nature et de la Vie, Université Mohamed Khider, 2024 |
| Format : | 1vol.(98p.) / ill., couv. ill. en coul / 30 cm |
| Langues: | Anglais |
| Mots-clés: | Digitization, E-Health, Artificial Intelligence, Machine Learning, Deep Learning, Paludism. |
| Résumé : |
In recent years, digitization has transformed medical practices globally, yet Algeria still faces significant healthcare challenges, including unequal access to specialized services, a shortage of medical professionals, and the compounded effects of an aging population and complex diseases. This thesis tackles two crucial challenges in this context by building a personalized AI-powered healthcare platform along with an effective diagnostic tool for Paludism. The proposed e-health platform aims to overcome existing barriers by providing a dedicated space for doctors and researchers. This platform features secure storage and access to a richrepository of healthcare-related academic papers, datasets, and models realized initially at Laboratoire de l’INFormatique Intelligente (LINFI) laboratory and eventually at other laboratories. Additionally, the current work introduces a robust Convolutional Neural Network (CNN) model designed for the precise classification of malaria-infected red blood cells. By employing various loss functions, several efficient techniques, and a hybrid dataset composed of images from a public dataset and images collected from seven cities in Algeria, with the assistance of paramedical specialists. Experimental results demonstrate the effectiveness of our approach. The custom CNN model achieved an accuracy of 99% with binary cross-entropy loss, and thus outperforming other tested models. The findings of this research promise substantial improvements in healthcare resource management and Paludism detection, ultimately contributing to better healthcare outcomes in Algeria and beyond. |
| Sommaire : |
Abstract I Résumé II Acknowledgements III List of Figures X List of Tables XI General Introduction 1 1 Medical Background 7 1.1 Introduction . . . . . . . . . . . . . . . . 7 1.2 HealthCare definition . . . . . . . . . . . . . . 8 1.3 Overview on healthcare practices . . . . . . . . . . . . . . . . . 8 1.3.1 Medical records management . . . . . . . . . . . . . . . . . 8 1.3.2 Diagnosis through blood analysis . . . . . . 8 1.3.3 Medical imaging diagnosis . . . . . . . . . . . . . . . . 9 1.4 Diagnosis based on medical imaging . . . . . . . . . . 9 1.4.1 Invasive imaging techniques . . . . . . . . . . . . . . . . . .. . 9 1.4.2 Non-invasive imaging techniques . . . . . . . . . . . . . . . . 1.5 Challenging issues related to medical diagnosis . . . . . . . . . . . . . 12 1.6 Healthcare situation in Algeria . . . . . . . . . . . . . . . . . . . . 13 1.7 E-Health platforms for medical assessment . . . . . . . . . . . 14 1.8 AI for E-Health . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . 15 1.9 Case study : Paludism Pathology . . . . . . . . . . . . . . . . . . . 18 1.9.1 Microscopic image overview . . . . . . . . . . . . . . . . . . 18 1.9.2 Paludism definition . . . . . . . . . . . . 19 1.9.3 Paludism life cycle . . . . . . . . . . . . . . . 19 1.9.4 Paludism diagnosis methods . . . . . . . . . . . . 20 1.10 Conclusion . . . . . . . . . . . . . . . . . . . 21 2 Theoretical and Technical Background 23 2.1 Introduction . . . . . . . . . . . . . . . . . . . 23 2.2 Artificial Intelligence . . . . . . . .. . . . . . . . . 23 2.3 Machine learning . . . . . . . . . . . . 24 2.4 Types of Learning . . . . . . . . . . . . . . . . . . . . . . 24 2.5 Deep Learning . . . . . . . . . . . . .. . 26 2.5.1 The differences between the Machine Learning and the Deep Learning . . . 27 2.5.2 Artificial Neural Networks . . . . . . . . . .. . 28 2.6 Convolutional Neural Network . . . . . . . . . . .. . . . . 34 2.6.1 CNN definition . . . . . . . . . . . . . . . . . . . . . . . . . 34 2.6.2 CNN Architecture . . . . . . . . . . . . . . . . . . . . . . 35 2.6.3 CNN layers configuration . . . . . . . . . . . . . . . . 39 2.6.4 Variation of CNN architectures . . . . . . . . . . . . 39 2.7 Related works of Paludism diagnosis model . . . . . . . . . . . 40 2.8 Conclusion . . . . . . . . . . . . . . . .. . . . . . 42 3 Towards AI-Powered Platform for E-Health 44 3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . 44 3.2 Requirements Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 3.2.1 Potential Users and Their Roles . . . . . . . 45 3.2.2 Principle Use Cases’ Descriptions . . . . . . . . .. . . . 47 3.2.3 Design considerations . . . . . . . . . .. . . . . 53 3.3 Platform Conception . . . . . . . . . . . . . . . . . . . . 54 3.3.1 Platform Architecture . . . . . . . . . . . . .. . . . . . 54 3.3.2 Sitemap . . . . . . . . . . . . . .. . . . . . . . 56 3.4 Platform Realization . . . 3.4.1 Used languages and tools: . . . . . . . . . . . . . . . . . 57 3.4.2 Interfaces’ screenshots . . . . . . . . . . . . . 58 4 Paludism Diagnosis Proposed Approach 68 4.1 Introduction . . . . . . . . . . . . . . . . .. . . . 68 4.2 General approach . . . . . . . . . . . . . . . . . . . . . . . . . 69 4.3 Dataset Description . . . . . . . . . . . . . . . . . . . . . 70 4.4 Image Data Preprocessing . . . . . . . . . . . . . . . . . . . . . . 70 4.5 Proposed approach . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . 71 4.6 Model requirements . . . . . . . . . . . . . . . . . . 72 4.6.1 Software requirements . . . . . . . . . 73 4.6.2 Hardware requirements . . . . . . . . . . . . . . . . 74 4.7 Training and Experimentation . . . . . . . . . . . . . . . . . . 74 4.8 Results and discussion . . . . . . . . . . . . . . . 75 4.9 Conclusion . . . . . . . . . . . . . . . . . . . . . 79 General Conclusion and Perspectives 80 References 82 |
| Type de document : | Mémoire master |
Disponibilité (1)
| Cote | Support | Localisation | Statut |
|---|---|---|---|
| MINF/892 | Mémoire master | bibliothèque sciences exactes | Consultable |
