Détection d’objet en temps réel en utilisant une approche basée sur l’apprentissage profond

General Information:

Cover Page:

Outline:

ACKNOWLEDGMENT:
CHAPTER 1: INTRODUCTION
1. INTRODUCTION:
1.1. Motivation and problem statement:
1.2. Scope of the master’s thesis:
1.3. Outline:
CHAPTER 2: BACKGROUND & RELATED WORK
2. BACKGROUND AND RELATED WORK:
2.1. Deep learning:
2.1.1. The history of deep Learning:
2.1.2. Types of learning:
2.1.2.1. Supervised Learning:
2.1.2.2. Unsupervised Learning:
2.1.2.3. Semi-supervised Learning:
2.1.3. Artificial neural network:
2.1.3.1. Multilayer perceptron(mlp):
2.1.3.2. Back-propagation:
2.1.3.3. Activation Function:
2.1.3.4. Batch Normalization:
2.1.4. Convolutional neural network:
2.1.4.1. The Convolutional Layer:
2.1.4.2. Pooling Layer:
2.1.4.3. Fully Connected Layer:
2.1.4.4. Why Convolutional Neural Networks?
2.2. Object detection:
2.2.1. History of detection algorithms:
2.2.2. Two-stage detection:
2.2.2.1. R-CNN:
2.2.2.2. Spp-net (Spatial Pyramid Pooling):
2.2.2.3. Fast R-CNN:
2.2.2.4. Faster R-CNN:
2.2.2.5. Mask R-CNN:
2.2.3. One-stage detection:
2.2.3.1. Yolo (you only look one):
2.2.3.1.1. IOU (Intersection over union):
2.2.3.1.2. Loss Function Explanations:
2.2.3.1.3. Non-max suppression:
2.2.3.1.4. Anchor boxes:
2.2.3.2. SSD (single shot detector):
2.2.3.3. Retina NET:
2.2.3.4. Yolo v2:
2.2.3.5. Yolo v3:
2.2.3.5.1. Backbone:
2.2.3.5.2. Feature Pyramids:
2.2.3.5.3. Loss function:
2.2.3.6. Tiny-yolov3:
2.2.3.7. Yolo v4:
2.2.3.7.1. Backbone:
2.2.3.7.2. SPP in YOLOv4:
2.2.3.7.3. Activation function:
2.2.3.7.4. Feature Pyramids:
2.2.3.8. Yolo v5:
2.2.3.7.5. Data Augmentation:
2.2.3.8.1. Backbone:
2.2.3.8.2. SPP:
2.2.3.8.3. Activation function:
2.2.3.8.4. Feature Pyramids:
2.2.3.8.5. Focus (also called by DepthToSpace):
2.2.4. Comparison of Faster-RCNN, YOLO, and SSD for Real-Time:
2.3. Conclusion:
CHAPTER 3: PROJECT DEVELOPMENT:
3. PROJECT DEVELOPMENT:
3.1. Performance metrics:
3.1.1.1. True Positive (TP):
3.1.1.2. True Negative (TN):
3.1.1.3. False Positive (FP):
3.1.1.4. False Negative (FN):
3.1.2. Average precision (AP):
3.1.3. Mean Average Precision (MAP):
3.1.4. Recall:
3.1.5. Precision:
3.2. Implementation:
3.2.1. Software Environment:
3.2.2. Hardware Environment:
3.2.3. Virtual environment:
3.2.4. Preparation of the data:
3.2.5. Object Detector:
3.2.6. Yolov3 architecture:
3.2.7. Training:
3.2.8. Results:
3.2.9. Conclusion:
GENERAL CONCLUSION:

Download The Thesis:

📥 تحميل الملف

For more
academic sources and references,
including theses and dissertations from Algerian universities,
, visit our main website.