Exploring Self- Supervised Vision Transformers for Medical Image Analysis (DINO-VIT) / Ikram Bentoumi

Public ISBD Titre : Exploring Self- Supervised Vision Transformers for Medical Image Analysis (DINO-VIT) Type de document : texte imprimé Auteurs : Ikram Bentoumi, Auteur ; Anfel Hibat-ellah Hamzaoui ; Moussaoui ,Abdelouahab, Directeur de thèse Editeur : Setif:UFA Année de publication : 2024 Importance : 1 vol (81 f .) Format : 29 cm Langues : Anglais (eng) Catégories : Thèses & Mémoires:Informatique Mots-clés : Deep Learning Computer Vision Medical Image Classification Vision Transformers (ViT) Convolutional Neural Networks (CNN), Index. décimale : 004 - Informatique Résumé : In the field of computer vision applied to the medical sector, convolutional neural networks (CNN) and recurrent neural networks (RNN) have long been the dominant techniques. However, other deep learning methods, such as Transformers, have gained popularity. This study compares the performance of several deep learning models, including Vision Transformers (ViT), on tasks involving the classification of brain and breast images.The results indicate that models based on EfficientNet and ResNet outperformed other architectures in terms of accuracy and error minimization. EfficientNet particularly excelled in the classification of brain images, while ResNet showed the best performance for breast images. Although ViTs maintained good Top-K accuracy, their overall performance was inferior compared to the other models. This analysis highlights the importance of selecting appropriate deep learning architectures for specific tasks to optimize results in medical image classification. Note de contenu : Sommaire General Introduction …………………………………………………………………………12 Problem and Objectives.……………………………………………………………………...12 Chapter 1: Theoritical Background ......................................................................................... 13 Introduction: ......................................................................................................................... 14 1.2 Artificial Intelligence (AI): ...................................................................................... 14 1.2.1 Goals of Artificial Intelligence: .......................................................................... 14 1.2.2 Types of Artificial Intelligence: ......................................................................... 14 1.2.3 Key AI Techniques: ........................................................................................... 15 1.2.4 Applications of Artificial Intelligence: .............................................................. 15 1.2.5 Challenges and Ethical Considerations of AI: ................................................... 15 1.2.6 Trends and Future of AI: .................................................................................... 15 1.3 Machine Learning (ML): .......................................................................................... 16 1.3.1 Types of Machine Learning: .............................................................................. 16 1.3.1.1 Supervised Learning: .................................................................................. 16 1.3.1.2 Unsupervised Learning: .............................................................................. 17 1.3.1.3 Semi-supervised Learning: ......................................................................... 17 1.3.1.4 Reinforcement Learning: ............................................................................ 18 1.3.2 Applications of Machine Learning: .................................................................... 18 1.4 Deep Learning: ......................................................................................................... 19 1.4.1 Deep learning models: ........................................................................................ 19 1.5 Self-Supervised Learning: ........................................................................................ 20 1.5.1 Key Concepts in Self-Supervised Learning: ...................................................... 20 1.5.2 Applications of Self-Supervised Learning: ........................................................ 21 1.6 Classification: ........................................................................................................... 21 1.6.1 Types of Classification Tasks: ........................................................................... 21 1.6.1.1 Number of Classes: ..................................................................................... 21 1.6.1.2 Nature of Data: ............................................................................................ 22 1.6.2 Applications of Classification in AI: .................................................................. 22 1.7 Conclusion: ............................................................................................................... 22 2 Chapter 2:Medical Imaging .............................................................................................. 24 2.1 Introduction: ............................................................................................................. 25 2.2 History of Medical Imaging: .................................................................................... 25 2.3 Medical imaging: ...................................................................................................... 26 2.3.1 Key Characteristics of Medical Imaging: ........................................................... 26 2.3.2 Description of Medical Images: ......................................................................... 26 2.3.3 Techniques of Medical Images: ......................................................................... 27 2.3.4 range of applications of Medical imaging: ......................................................... 31 2.3.5 Importance of Medical Imaging: ........................................................................ 31 2.3.6 Challenges and Considerations in Medical Imaging: ......................................... 32 2.4 Conclusion: ............................................................................................................... 32 3 Chapter3: Vision Transformers (DINO-VIT) ................................................................... 33 3.1 Introduction: ............................................................................................................. 34 3.2 Vision Transformers: ................................................................................................ 34 3.3 How Vision Transformers works ? .......................................................................... 34 3.3.1 Transformer Foundation: .................................................................................... 34 3.3.2 Adapting the Transformer for Images: ............................................................... 35 3.3.3 Advantages of Vision Transformers in Medical Imaging: ................................. 35 3.4 DINo-VIT Method: .................................................................................................. 36 3.5 Advantages of DINo-VIT Method: .......................................................................... 36 3.6 Applications of Vision Transformers and DINo-VIT Method: ............................... 36 3.7 Conclusion: ............................................................................................................... 37 4 Chapter 4: Methodology and experiments. ....................................................................... 38 4.1 Introduction: ............................................................................................................. 39 4.2 Dataset and Preprocessing: ....................................................................................... 39 4.2.1 Dataset Description: ........................................................................................... 39 4.2.1.1 Dataset1: Brain cancer. ............................................................................... 39 4.2.1.2 Dataset02: BreastTumor. ............................................................................ 41 4.3 Development environment: ...................................................................................... 42 4.3.1 Software: ............................................................................................................ 43 4.3.2 Hardware: ........................................................................................................... 44 4.4 Methodology and experiments results:..................................................................... 44 4.4.1 Proposed Methodes: ........................................................................................... 44 4.4.1.1 Dino ViT (Vision Transformer): ................................................................. 44 4.4.2 Explanation: ....................................................................................................... 44 4.4.2.1 Convolutional Neural Network (CNN): ...................................................... 46 4.4.2.2 ResNet50: .................................................................................................... 47 4.4.2.3 EfficientNetB0 : .......................................................................................... 49 4.4.3 Experiments Results: .......................................................................................... 51 4.4.3.1 Vision Transformer (ViT): .......................................................................... 51 4.4.3.1.1 Dataset1: Brain Cancer. .......................................................................... 51 4.4.3.1.2 Dataset02: Breast Tumor. ....................................................................... 54 4.4.3.2 Convolutional Neural Network (CNN) : .................................................... 56 4.4.3.2.1 Dataset01: Brain Cancer. ........................................................................ 56 4.4.3.2.2 Dataset02: Breast Tumor. ....................................................................... 58 4.4.3.3 ResNet50: .................................................................................................... 61 4.4.3.3.1 Dataset01 : Brain Cancer. ....................................................................... 61 4.4.3.3.2 Dataset02 : Breast Tumor. ...................................................................... 63 4.4.3.4 EfficientB0: ................................................................................................. 66 4.4.3.4.1 Dataset01: Brain Cancer. ........................................................................ 66 4.4.3.4.2 Dataset02 : Breast Tumor. ...................................................................... 69 5 Chapter 5: Discussion. ...................................................................................................... 73 5.1 Analysis and Comparison: ....................................................................................... 74 5.1.1 Dataset 01: Brain tumor. .................................................................................... 74 5.1.1.1 Loss: ............................................................................................................ 74 5.1.1.2 Accuracy: .................................................................................................... 74 5.1.1.3 Top-k Accuracy: ......................................................................................... 74 5.1.1.4 Conclusion: ................................................................................................. 74 5.1.2 Dataset 02: Breast tumor. ................................................................................... 75 5.1.2.1 Loss: ............................................................................................................ 75 5.1.2.2 Accuracy: .................................................................................................... 75 5.1.2.3 Top-K Accuracy: ........................................................................................ 75 5.1.2.4 Conclusion: ................................................................................................. 75 5.2 Conclusion: ............................................................................................................... 76 General Conclusion…………………………………………………………………………...77 Côte titre : MAI/0912 Exploring Self- Supervised Vision Transformers for Medical Image Analysis (DINO-VIT) [texte imprimé] / Ikram Bentoumi, Auteur ; Anfel Hibat-ellah Hamzaoui ; Moussaoui ,Abdelouahab, Directeur de thèse . - [S.l.] : Setif:UFA, 2024 . - 1 vol (81 f .) ; 29 cm. Langues : Anglais (eng) Catégories : Thèses & Mémoires:Informatique Mots-clés : Deep Learning Computer Vision Medical Image Classification Vision Transformers (ViT) Convolutional Neural Networks (CNN), Index. décimale : 004 - Informatique Résumé : In the field of computer vision applied to the medical sector, convolutional neural networks (CNN) and recurrent neural networks (RNN) have long been the dominant techniques. However, other deep learning methods, such as Transformers, have gained popularity. This study compares the performance of several deep learning models, including Vision Transformers (ViT), on tasks involving the classification of brain and breast images.The results indicate that models based on EfficientNet and ResNet outperformed other architectures in terms of accuracy and error minimization. EfficientNet particularly excelled in the classification of brain images, while ResNet showed the best performance for breast images. Although ViTs maintained good Top-K accuracy, their overall performance was inferior compared to the other models. This analysis highlights the importance of selecting appropriate deep learning architectures for specific tasks to optimize results in medical image classification. Note de contenu : Sommaire General Introduction …………………………………………………………………………12 Problem and Objectives.……………………………………………………………………...12 Chapter 1: Theoritical Background ......................................................................................... 13 Introduction: ......................................................................................................................... 14 1.2 Artificial Intelligence (AI): ...................................................................................... 14 1.2.1 Goals of Artificial Intelligence: .......................................................................... 14 1.2.2 Types of Artificial Intelligence: ......................................................................... 14 1.2.3 Key AI Techniques: ........................................................................................... 15 1.2.4 Applications of Artificial Intelligence: .............................................................. 15 1.2.5 Challenges and Ethical Considerations of AI: ................................................... 15 1.2.6 Trends and Future of AI: .................................................................................... 15 1.3 Machine Learning (ML): .......................................................................................... 16 1.3.1 Types of Machine Learning: .............................................................................. 16 1.3.1.1 Supervised Learning: .................................................................................. 16 1.3.1.2 Unsupervised Learning: .............................................................................. 17 1.3.1.3 Semi-supervised Learning: ......................................................................... 17 1.3.1.4 Reinforcement Learning: ............................................................................ 18 1.3.2 Applications of Machine Learning: .................................................................... 18 1.4 Deep Learning: ......................................................................................................... 19 1.4.1 Deep learning models: ........................................................................................ 19 1.5 Self-Supervised Learning: ........................................................................................ 20 1.5.1 Key Concepts in Self-Supervised Learning: ...................................................... 20 1.5.2 Applications of Self-Supervised Learning: ........................................................ 21 1.6 Classification: ........................................................................................................... 21 1.6.1 Types of Classification Tasks: ........................................................................... 21 1.6.1.1 Number of Classes: ..................................................................................... 21 1.6.1.2 Nature of Data: ............................................................................................ 22 1.6.2 Applications of Classification in AI: .................................................................. 22 1.7 Conclusion: ............................................................................................................... 22 2 Chapter 2:Medical Imaging .............................................................................................. 24 2.1 Introduction: ............................................................................................................. 25 2.2 History of Medical Imaging: .................................................................................... 25 2.3 Medical imaging: ...................................................................................................... 26 2.3.1 Key Characteristics of Medical Imaging: ........................................................... 26 2.3.2 Description of Medical Images: ......................................................................... 26 2.3.3 Techniques of Medical Images: ......................................................................... 27 2.3.4 range of applications of Medical imaging: ......................................................... 31 2.3.5 Importance of Medical Imaging: ........................................................................ 31 2.3.6 Challenges and Considerations in Medical Imaging: ......................................... 32 2.4 Conclusion: ............................................................................................................... 32 3 Chapter3: Vision Transformers (DINO-VIT) ................................................................... 33 3.1 Introduction: ............................................................................................................. 34 3.2 Vision Transformers: ................................................................................................ 34 3.3 How Vision Transformers works ? .......................................................................... 34 3.3.1 Transformer Foundation: .................................................................................... 34 3.3.2 Adapting the Transformer for Images: ............................................................... 35 3.3.3 Advantages of Vision Transformers in Medical Imaging: ................................. 35 3.4 DINo-VIT Method: .................................................................................................. 36 3.5 Advantages of DINo-VIT Method: .......................................................................... 36 3.6 Applications of Vision Transformers and DINo-VIT Method: ............................... 36 3.7 Conclusion: ............................................................................................................... 37 4 Chapter 4: Methodology and experiments. ....................................................................... 38 4.1 Introduction: ............................................................................................................. 39 4.2 Dataset and Preprocessing: ....................................................................................... 39 4.2.1 Dataset Description: ........................................................................................... 39 4.2.1.1 Dataset1: Brain cancer. ............................................................................... 39 4.2.1.2 Dataset02: BreastTumor. ............................................................................ 41 4.3 Development environment: ...................................................................................... 42 4.3.1 Software: ............................................................................................................ 43 4.3.2 Hardware: ........................................................................................................... 44 4.4 Methodology and experiments results:..................................................................... 44 4.4.1 Proposed Methodes: ........................................................................................... 44 4.4.1.1 Dino ViT (Vision Transformer): ................................................................. 44 4.4.2 Explanation: ....................................................................................................... 44 4.4.2.1 Convolutional Neural Network (CNN): ...................................................... 46 4.4.2.2 ResNet50: .................................................................................................... 47 4.4.2.3 EfficientNetB0 : .......................................................................................... 49 4.4.3 Experiments Results: .......................................................................................... 51 4.4.3.1 Vision Transformer (ViT): .......................................................................... 51 4.4.3.1.1 Dataset1: Brain Cancer. .......................................................................... 51 4.4.3.1.2 Dataset02: Breast Tumor. ....................................................................... 54 4.4.3.2 Convolutional Neural Network (CNN) : .................................................... 56 4.4.3.2.1 Dataset01: Brain Cancer. ........................................................................ 56 4.4.3.2.2 Dataset02: Breast Tumor. ....................................................................... 58 4.4.3.3 ResNet50: .................................................................................................... 61 4.4.3.3.1 Dataset01 : Brain Cancer. ....................................................................... 61 4.4.3.3.2 Dataset02 : Breast Tumor. ...................................................................... 63 4.4.3.4 EfficientB0: ................................................................................................. 66 4.4.3.4.1 Dataset01: Brain Cancer. ........................................................................ 66 4.4.3.4.2 Dataset02 : Breast Tumor. ...................................................................... 69 5 Chapter 5: Discussion. ...................................................................................................... 73 5.1 Analysis and Comparison: ....................................................................................... 74 5.1.1 Dataset 01: Brain tumor. .................................................................................... 74 5.1.1.1 Loss: ............................................................................................................ 74 5.1.1.2 Accuracy: .................................................................................................... 74 5.1.1.3 Top-k Accuracy: ......................................................................................... 74 5.1.1.4 Conclusion: ................................................................................................. 74 5.1.2 Dataset 02: Breast tumor. ................................................................................... 75 5.1.2.1 Loss: ............................................................................................................ 75 5.1.2.2 Accuracy: .................................................................................................... 75 5.1.2.3 Top-K Accuracy: ........................................................................................ 75 5.1.2.4 Conclusion: ................................................................................................. 75 5.2 Conclusion: ............................................................................................................... 76 General Conclusion…………………………………………………………………………...77 Côte titre : MAI/0912

Exemplaires (1)

Code-barres	Cote	Support	Localisation	Section	Disponibilité
MAI/0912	MAI/0912	Mémoire	Bibliothéque des sciences	Anglais	Disponible Disponible

---

1 (1 - 1 / 1)