University Sétif 1 FERHAT ABBAS Faculty of Sciences
Détail de l'auteur
Auteur Mohamed Yacine Haddad |
Documents disponibles écrits par cet auteur
Ajouter le résultat dans votre panier Affiner la recherche
Titre : Automatic spice classification application by image recognition Type de document : document électronique Auteurs : Maamar Karfa ; Mohamed Yacine Haddad, Auteur ; Mekroud,Noureddine, Directeur de thèse Editeur : Setif:UFA Année de publication : 2025 Importance : 1 vol (63 f .) Format : 29 cm Langues : Anglais (eng) Catégories : Thèses & Mémoires:Informatique Mots-clés : Shape recognition
Deep learning
Spice classification
EfficientNetB3Index. décimale : 004 Informatique Résumé :
This thesis investigates the integration of shape recognition and deep learning techniques to
develop a mobile-based spice classification system. Addressing the challenges of visual
similarity and texture variability among spices, the research utilizes EfficientNetB3 as a
transfer learning backbone, achieving a classification accuracy of 93.78%. The system
combines Flutter for cross-platform frontend development with Django for backend
processing, ensuring real-time performance and scalability. Key contributions include a curated
dataset of 24 spice categories, optimized preprocessing pipelines, and mobile-specific model
compression techniques. The thesis also discusses ethical considerations in AI deployment and
potential future enhancements, such as augmented reality guidance. The results demonstrate
that shape recognition can be effectively adapted for culinary applications, establishing a
benchmark for mobile computer vision systems.Note de contenu : Sommaire
Chapter 1: Shape Recognition
Theoretical Background
1 Introduction 1
2 Shape Recognition 1
2.1 Basic Theories of Shape Recognition ------------------------------------------------------ 1
2.1.1 Pattern Recognition Theory 1
2.1.2 Cognitive Theory 1
2.2 Techniques for Shape Recognition 2
2.2.1 Traditional Methods 2
2.2.2 Machine Learning-Based Methods 3
2.3 Explainability and Interpretability in Image Recognition ---------------------------------- 3
2.4 Common Applications of Shape Recognition --------------------------------------------- 4
2.4.1 Introduction to Applications of Shape Recognition ------------------------------------ 4
3 Image Recognition in Mobile Applications -------------------------------------------------------- 7
3.1 Core Technologies Behind Image Recognition -------------------------------------------- 7
3.2 Popular Mobile Image Recognition Frameworks ------------------------------------------ 7
3.3 Limitations of Image Recognition in Mobile Apps ----------------------------------------- 8
3.4 Ethical Considerations in AI-based Food Classification ----------------------------------- 8
3.5 Future Prospects of Image Recognition in Mobile Applications -------------------------- 8
4 Spice Specification 9
4.1 Importance of Spice Classification 9
4.2 Challenges in Spice Classification 10
4.3 Traditional Methods of Spice Classification ----------------------------------------------- 10
4.4 Modern Techniques for Spice Classification ---------------------------------------------- 11
4.5 Previous Studies and Research on Spice Classification---------------------------------- 11
5 Conclusion 11
Chapter 2 : Application Design
1 Introduction 13
2 Needs Analysis 13
2.1 Identification of target users 13
2.2 Methods for Identifying Target Users ------------------------------------------------------ 13
2.3 Functional and non-functional specifications -------------------------------------------- 13
2.3.1 Functional Specifications 13
2.3.2 Non-Functional Specifications 14
2.4 Security and Data Privacy Architecture --------------------------------------------------- 14
3 Application Architecture 14
3.1 Technological Choices 15
3.2 Development Tools 15
3.2.1 IDE: Visual Studio Code (VS Code) 15
3.2.2 Dataset Management: Kaggle 15
3.2.3 Machine Learning Training 16
3.2.4 Local Testing C Deployment 16
3.3 Layered Architecture Model 17
3.3.1 Presentation Layer (Frontend - Flutter) ------------------------------------------------ 17
3.3.2 Business Logic Layer (Backend - Django REST Framework) ------------------------- 17
3.3.3 Data Processing Layer 18
3.4 Equivalent Works 18
3.5 Contribution 19
3.5.1 Proposal Platform (Spicyfy) 19
3.5.2 Comparison Between Our Approach and Equivalent Approaches ------------------- 19
3.6 Transfer Learning in Deep Learning 19
3.6.1 How Transfer Learning Works with CNNs ---------------------------------------------- 20
3.6.2 EfficientNet Architecture 20
3.6.3 Comparison of CNN Architectures ----------------------------------------------------- 21
4 System Design 23
4.1 UML Diagrams 23
4.1.1 Class Diagrams 23
4.1.2 Sequence Diagrams 24
4.1.3 Component Diagrams 24
User Interface 25
4.2 Principles of UI/UX Design 25
4.3 Visual Design Principles 25
4.4 Prototyping and Iterations 26
4.4.1 Prototyping Approach 26
5 Technical Specifications of Spice Identification App ------------------------------------------- 27
6 Conclusion 30
Chapter 3:Application Development
1 Introduction 32
2 Frontend Development 32
2.1 Languages and Frameworks Used 32
2.1.1 Development Environment 32
2.1.2 State Management 33
2.1.3 UI Components 34
2.1.4 Essential Packages 34
2.2 Implementation of main features 35
2.2.1 Core Feature Architecture 35
2.2.2 Key Feature Breakdown 35
2.2.3 Technical Implementation 37
2.2.4 Cross-Feature Integration 37
2.2.5 Quality Assurance 38
2.3 App Screen Architecture 38
2.3.1 Screen Hierarchy C Navigation Flow --------------------------------------------------- 38
3 Backend Development 42
3.1 Server and database configuration 42
3.2 APIs and web services 42
4 Shape Recognition Integration 43
4.1 Dataset Curation and Preparation 43
4.1.1 Dataset Composition 43
4.1.2 Class Distribution and Challenges 45
4.1.3 Data Collection Methodology 45
4.2 Algorithms and techniques used 45
4.3 Model Implementation with Transfer Learning ------------------------------------------- 46
4.3.1 TensorFlow C Keras 46
4.3.2 EfficientNetB3 (Pre-trained CNN Backbone) ------------------------------------------ 47
4.3.3 OpenCV (via TensorFlow + Keras) 47
4.3.4 Matplotlib 47
4.3.5 Additional Libraries 47
4.4 Model training and optimization 47
4.4.1 Dataset Splitting Strategy 48
4.4.2 Training Accuracy Curve Analysis 49
4.4.3 Choice of hyperparameters for the basic architecture-------------------------------- 50
4.5 Performance Evaluation and Continuous Improvement -------------------------------- 56
5 Conclusion 58Côte titre : MAI/1015 Automatic spice classification application by image recognition [document électronique] / Maamar Karfa ; Mohamed Yacine Haddad, Auteur ; Mekroud,Noureddine, Directeur de thèse . - [S.l.] : Setif:UFA, 2025 . - 1 vol (63 f .) ; 29 cm.
Langues : Anglais (eng)
Catégories : Thèses & Mémoires:Informatique Mots-clés : Shape recognition
Deep learning
Spice classification
EfficientNetB3Index. décimale : 004 Informatique Résumé :
This thesis investigates the integration of shape recognition and deep learning techniques to
develop a mobile-based spice classification system. Addressing the challenges of visual
similarity and texture variability among spices, the research utilizes EfficientNetB3 as a
transfer learning backbone, achieving a classification accuracy of 93.78%. The system
combines Flutter for cross-platform frontend development with Django for backend
processing, ensuring real-time performance and scalability. Key contributions include a curated
dataset of 24 spice categories, optimized preprocessing pipelines, and mobile-specific model
compression techniques. The thesis also discusses ethical considerations in AI deployment and
potential future enhancements, such as augmented reality guidance. The results demonstrate
that shape recognition can be effectively adapted for culinary applications, establishing a
benchmark for mobile computer vision systems.Note de contenu : Sommaire
Chapter 1: Shape Recognition
Theoretical Background
1 Introduction 1
2 Shape Recognition 1
2.1 Basic Theories of Shape Recognition ------------------------------------------------------ 1
2.1.1 Pattern Recognition Theory 1
2.1.2 Cognitive Theory 1
2.2 Techniques for Shape Recognition 2
2.2.1 Traditional Methods 2
2.2.2 Machine Learning-Based Methods 3
2.3 Explainability and Interpretability in Image Recognition ---------------------------------- 3
2.4 Common Applications of Shape Recognition --------------------------------------------- 4
2.4.1 Introduction to Applications of Shape Recognition ------------------------------------ 4
3 Image Recognition in Mobile Applications -------------------------------------------------------- 7
3.1 Core Technologies Behind Image Recognition -------------------------------------------- 7
3.2 Popular Mobile Image Recognition Frameworks ------------------------------------------ 7
3.3 Limitations of Image Recognition in Mobile Apps ----------------------------------------- 8
3.4 Ethical Considerations in AI-based Food Classification ----------------------------------- 8
3.5 Future Prospects of Image Recognition in Mobile Applications -------------------------- 8
4 Spice Specification 9
4.1 Importance of Spice Classification 9
4.2 Challenges in Spice Classification 10
4.3 Traditional Methods of Spice Classification ----------------------------------------------- 10
4.4 Modern Techniques for Spice Classification ---------------------------------------------- 11
4.5 Previous Studies and Research on Spice Classification---------------------------------- 11
5 Conclusion 11
Chapter 2 : Application Design
1 Introduction 13
2 Needs Analysis 13
2.1 Identification of target users 13
2.2 Methods for Identifying Target Users ------------------------------------------------------ 13
2.3 Functional and non-functional specifications -------------------------------------------- 13
2.3.1 Functional Specifications 13
2.3.2 Non-Functional Specifications 14
2.4 Security and Data Privacy Architecture --------------------------------------------------- 14
3 Application Architecture 14
3.1 Technological Choices 15
3.2 Development Tools 15
3.2.1 IDE: Visual Studio Code (VS Code) 15
3.2.2 Dataset Management: Kaggle 15
3.2.3 Machine Learning Training 16
3.2.4 Local Testing C Deployment 16
3.3 Layered Architecture Model 17
3.3.1 Presentation Layer (Frontend - Flutter) ------------------------------------------------ 17
3.3.2 Business Logic Layer (Backend - Django REST Framework) ------------------------- 17
3.3.3 Data Processing Layer 18
3.4 Equivalent Works 18
3.5 Contribution 19
3.5.1 Proposal Platform (Spicyfy) 19
3.5.2 Comparison Between Our Approach and Equivalent Approaches ------------------- 19
3.6 Transfer Learning in Deep Learning 19
3.6.1 How Transfer Learning Works with CNNs ---------------------------------------------- 20
3.6.2 EfficientNet Architecture 20
3.6.3 Comparison of CNN Architectures ----------------------------------------------------- 21
4 System Design 23
4.1 UML Diagrams 23
4.1.1 Class Diagrams 23
4.1.2 Sequence Diagrams 24
4.1.3 Component Diagrams 24
User Interface 25
4.2 Principles of UI/UX Design 25
4.3 Visual Design Principles 25
4.4 Prototyping and Iterations 26
4.4.1 Prototyping Approach 26
5 Technical Specifications of Spice Identification App ------------------------------------------- 27
6 Conclusion 30
Chapter 3:Application Development
1 Introduction 32
2 Frontend Development 32
2.1 Languages and Frameworks Used 32
2.1.1 Development Environment 32
2.1.2 State Management 33
2.1.3 UI Components 34
2.1.4 Essential Packages 34
2.2 Implementation of main features 35
2.2.1 Core Feature Architecture 35
2.2.2 Key Feature Breakdown 35
2.2.3 Technical Implementation 37
2.2.4 Cross-Feature Integration 37
2.2.5 Quality Assurance 38
2.3 App Screen Architecture 38
2.3.1 Screen Hierarchy C Navigation Flow --------------------------------------------------- 38
3 Backend Development 42
3.1 Server and database configuration 42
3.2 APIs and web services 42
4 Shape Recognition Integration 43
4.1 Dataset Curation and Preparation 43
4.1.1 Dataset Composition 43
4.1.2 Class Distribution and Challenges 45
4.1.3 Data Collection Methodology 45
4.2 Algorithms and techniques used 45
4.3 Model Implementation with Transfer Learning ------------------------------------------- 46
4.3.1 TensorFlow C Keras 46
4.3.2 EfficientNetB3 (Pre-trained CNN Backbone) ------------------------------------------ 47
4.3.3 OpenCV (via TensorFlow + Keras) 47
4.3.4 Matplotlib 47
4.3.5 Additional Libraries 47
4.4 Model training and optimization 47
4.4.1 Dataset Splitting Strategy 48
4.4.2 Training Accuracy Curve Analysis 49
4.4.3 Choice of hyperparameters for the basic architecture-------------------------------- 50
4.5 Performance Evaluation and Continuous Improvement -------------------------------- 56
5 Conclusion 58Côte titre : MAI/1015 Exemplaires (1)
Code-barres Cote Support Localisation Section Disponibilité MAI/1015 MAI/1015 Mémoire Bibliothèque des sciences Anglais Disponible
Disponible
