List of Figures

1. Different viewpoints of the same breast.
2. Shapes and margins of a mass.
3. Examples of masses.
4. HRIMAC's architecture.
5. Scheme of our proposal.
6. Breast profile segmentation using the fast breast segmentation.
7. Breast profile segmentation using the contour-based approach.
8. Right and left mammograms of a woman.
9. CC and MLO views of the same breast.
10. A tumour-like template.
11. FROC analysis of the algorithms over the set of $120$ mammograms.
12. Grey-level histograms of a breast.
13. Mammogram division strategies for feature extraction.
14. Quantitatively reviewed strategies for breast segmentation.
15. Template models classification.
16. Four RoIs and their manual segmentations.
17. Eigenmasses.
18. Constructed templates for a posterior pattern matching.
19. Profile of the constructed templates.
20. Potential images of three mammograms.
21. Suspicious regions found using the Bayesian pattern matching approach.
22. FROC analysis.
23. FROC analysis per lesion size.
24. Mean $A_z$ values for the MIAS database.
25. Mean Kappa values for the MIAS database.
26. Mean $A_z$ values for the DDSM database.
27. Mean Kappa values for the DDSM database.
28. Eigenrois.
29. FROC analysis of the proposals.
30. FROC analysis comparison.
31. FROC analysis comparison per lesion size.
32. Influence of the training database in the FROC analysis.
33. FROC analysis comparison using different databases.
34. Proposed algorithms FROC comparison.
35. Comparison of the proposal with other algorithms.
36. Mean $A_z$ value for the DDSM database.
37. Mean Kappa values for the DDSM database.
38. Breast tissue information.
39. Histogram of a typical mammogram.
40. Sequence of the breast profile segmentation.
41. Examples of the breast profile segmentation.
42. Contour growing scheme.
43. Cost functions for robust candidate selection.
44. Two examples of the breast-skin line segmentation.
45. ROC example.