Chapter 7 Target Detection

1. 7.1 Candidate-based target detector 228

2. 7.1.1 Sliding Window Detector 228

3. 7.1.2 Selective Search 229

4. 7.1.3 R-CNN 230

5. 7.1.4 Boundary Box Regressor 230

6. 7.1.5 Fast R-CNN 231

7. 7.1.6 ROI Pooling 233

8. 7.1.7 Faster R-CNN 233

9. 7.1.8 Candidate Area Network 234

10. 7.1.9 Performance of the R-CNN method 236

11. 7.2 Area-based full convolutional neural network (R-FCN) 237

12. 7.3 Single Target Detector 240

13. 7.3.1 Single detector 241

14. 7.3.2 Sliding window for prediction 241

15. 7.3.3 SSD 243

16. 7.4 YOLO Series 244

17. 7.4.1 Introduction to YOLOv1 244

18. 7.4.2 What are the advantages and disadvantages of the YOLOv1 model? 252

19. 7.4.3 YOLOv2 253

20. 7.4.4 YOLOv2 Improvement Strategy 254

21. 7.4.5 Training of YOLOv2 261

22. 7.4.6 YOLO9000 261

23. 7.4.7 YOLOv3 263

24. 7.4.8 YOLOv3 Improvements 264

P