Machine Learning in MRI: From Methods to Clinical Translation, Volume Thirteen in the
Advances in Magnetic Resonance Technology and Applications series presents state-of-the-art machine learning methods in magnetic resonance imaging that can shape and impact the future of patient treatment and planning. Common methods and strategies along the processing chain of data acquisition, image reconstruction, image post-processing, and image analysis of these imaging modalities are presented and illustrated. The book focuses on applications and anatomies for which machine learning methods can bring, or have already brought. Ideas and concepts on how processing could be harmonized and used to provide deployable frameworks that integrate into the clinical workflows are also considered.
Pitfalls and current limitations are discussed in the context of how they could be overcome to cater for clinical needs, making this an ideal reference for medical imaging researchers, industry scientists and engineers, advanced undergraduate and graduate students, and clinicians. By giving an interdisciplinary presentation and discussion on the obstacles and possible solutions for the clinical translation of machine learning methods, this book enables the evolution of machine learning in medical imaging for the next decade.
1. 1.1 Introduction - p. 3 2. 1.2 Vector and matrix calculation - p. 3 3. 1.3 Random variable and probability distribution - p. 6 4. 1.4 Kullback-Leibler divergence and cross entropy - p. 10 5. 1.5 Probabilistic framework of machine learning - p. 12 6. 1.6 Estimators - p. 19 7. 1.7 Markov and hidden Markov process - p. 21 8. Chapter 2. The ingredients for machine learning - p. 23 9. 2.1 Introduction - p. 23 10. 2.2 Organization of this chapter - p. 24 11. 2.3 Learning models - p. 25 12. 2.4 Types of learning - p. 25 13. 2.5 Cost function, optimization, and backpropagation - p. 26 14. 2.6 Database splitting - p. 30 15. 2.7 Fundamental network elements - p. 32 16. 2.8 Fundamental architectures - p. 45 17. 2.9 Generative models - p. 50 18. 2.10 Summary - p. 54 19. 2.11 Further resources and tutorials - p. 55 20. Chapter 3. Introduction to the physics behind MR - p. 65 21. Part 2 MR image acquisition - p. P2 22. Chapter 4. Adjust to your imaging scenario: Learning and optimizing MRI data sampling - p. 91 23. Chapter 5. MR imaging in the low-field: Leveraging the power of machine learning - p. 129 24. Chapter 6. The smart spin: Machine learning for magnetic resonance spectroscopy - p. 165 25. Part 3 MR image reconstruction - p. P3 26. Chapter 7. Get the image: Machine learning for MR image reconstruction - p. 189 27. Chapter 8. Enhance the image: Super resolution in MRI - p. 219 28. Chapter 9. Freeze the motion: Machine learning for motion correction - p. 249 29. Chapter 10. Map the image: Machine learning for quantitative MR mapping - p. 275 30. Chapter 11. Am (A)I hallucinating? Nonrobustness, hallucinations, and unpredictable performance of AI for MR image reconstruction - p. 303 31. Part 4 MR image postprocessing - p. P4 32. Chapter 12. Cut it here: Image segmentation for MRI - p. 329 33. 12.1 Introduction - p. 329 34. 12.2 Traditional segmentation methods - p. 330 35. 12.3 Deep learning-based segmentation methods - p. 333 36. 12.4 Neutral network training and losses - p. 342 37. 12.5 Evaluation metrics for semantic segmentation - p. 351 38. 12.6 Challenges of existing deep learning-based methods and strategies to alleviate - p. 355 39. 12.7 Clinical applications of semantic segmentation in MR imaging - p. 361 40. 12.8 Future directions - p. 369 41. 12.9 Open-source tools for medical image segmentation - p. 370 42. 12.10 Summary - p. 372 43. Chapter 13. Quality matters: Automated MR image quality control - p. 389 44. Chapter 14. What is beyond the image? Machine learning for MR image analysis - p. 417 45. Chapter 15. Give me that other image: Machine learning for image-to-image translation - p. 439 46. Part 5 Generalization and fairness - p. P5 47. Chapter 16. The cause and effect of an MR image: Robustness and generalizability - p. 465 48. Chapter 17. Scale it up: Large-scale MR data processing - p. 487 49. Chapter 18. Human in the loop: Integration of experts in MR data processing - p. 509 50. Part 6 Clinical application - p. P6 51. Chapter 19. Machine learning applications in neuroradiology - p. 545 52. Chapter 20. AI in cardiovascular MRI - p. 565 53. Chapter 21. Machine learning in body MRI - p. 589 54. Chapter 22. Clinical applications of machine learning in breast MRI - p. 609 55. Chapter 23. Clinical applications of machine learning in musculoskeletal MRI - p. 621 56. Part 7 Reproducibility - p. P7 57. Chapter 24. Let's share: Open-source frameworks and public databases - p. 643 58. Chapter 25. System under test: Challenges for algorithm benchmarking - p. 667 59. Part 8 Conclusion - p. P8 60. Chapter 26. Current solutions, future challenges, and directions - p. 687 61. Index - p. 697
Machine Learning in MRI: From Methods to Clinical Translation, Volume Thirteen in the
Advances in Magnetic Resonance Technology and Applications series presents state-of-the-art machine learning methods in magnetic resonance imaging that can shape and impact the future of patient treatment and planning. Common methods and strategies along the processing chain of data acquisition, image reconstruction, image post-processing, and image analysis of these imaging modalities are presented and illustrated. The book focuses on applications and anatomies for which machine learning methods can bring, or have already brought. Ideas and concepts on how processing could be harmonized and used to provide deployable frameworks that integrate into the clinical workflows are also considered.
Pitfalls and current limitations are discussed in the context of how they could be overcome to cater for clinical needs, making this an ideal reference for medical imaging researchers, industry scientists and engineers, advanced undergraduate and graduate students, and clinicians. By giving an interdisciplinary presentation and discussion on the obstacles and possible solutions for the clinical translation of machine learning methods, this book enables the evolution of machine learning in medical imaging for the next decade.
1. 1.1 Introduction - p. 3 2. 1.2 Vector and matrix calculation - p. 3 3. 1.3 Random variable and probability distribution - p. 6 4. 1.4 Kullback-Leibler divergence and cross entropy - p. 10 5. 1.5 Probabilistic framework of machine learning - p. 12 6. 1.6 Estimators - p. 19 7. 1.7 Markov and hidden Markov process - p. 21 8. Chapter 2. The ingredients for machine learning - p. 23 9. 2.1 Introduction - p. 23 10. 2.2 Organization of this chapter - p. 24 11. 2.3 Learning models - p. 25 12. 2.4 Types of learning - p. 25 13. 2.5 Cost function, optimization, and backpropagation - p. 26 14. 2.6 Database splitting - p. 30 15. 2.7 Fundamental network elements - p. 32 16. 2.8 Fundamental architectures - p. 45 17. 2.9 Generative models - p. 50 18. 2.10 Summary - p. 54 19. 2.11 Further resources and tutorials - p. 55 20. Chapter 3. Introduction to the physics behind MR - p. 65 21. Part 2 MR image acquisition - p. P2 22. Chapter 4. Adjust to your imaging scenario: Learning and optimizing MRI data sampling - p. 91 23. Chapter 5. MR imaging in the low-field: Leveraging the power of machine learning - p. 129 24. Chapter 6. The smart spin: Machine learning for magnetic resonance spectroscopy - p. 165 25. Part 3 MR image reconstruction - p. P3 26. Chapter 7. Get the image: Machine learning for MR image reconstruction - p. 189 27. Chapter 8. Enhance the image: Super resolution in MRI - p. 219 28. Chapter 9. Freeze the motion: Machine learning for motion correction - p. 249 29. Chapter 10. Map the image: Machine learning for quantitative MR mapping - p. 275 30. Chapter 11. Am (A)I hallucinating? Nonrobustness, hallucinations, and unpredictable performance of AI for MR image reconstruction - p. 303 31. Part 4 MR image postprocessing - p. P4 32. Chapter 12. Cut it here: Image segmentation for MRI - p. 329 33. 12.1 Introduction - p. 329 34. 12.2 Traditional segmentation methods - p. 330 35. 12.3 Deep learning-based segmentation methods - p. 333 36. 12.4 Neutral network training and losses - p. 342 37. 12.5 Evaluation metrics for semantic segmentation - p. 351 38. 12.6 Challenges of existing deep learning-based methods and strategies to alleviate - p. 355 39. 12.7 Clinical applications of semantic segmentation in MR imaging - p. 361 40. 12.8 Future directions - p. 369 41. 12.9 Open-source tools for medical image segmentation - p. 370 42. 12.10 Summary - p. 372 43. Chapter 13. Quality matters: Automated MR image quality control - p. 389 44. Chapter 14. What is beyond the image? Machine learning for MR image analysis - p. 417 45. Chapter 15. Give me that other image: Machine learning for image-to-image translation - p. 439 46. Part 5 Generalization and fairness - p. P5 47. Chapter 16. The cause and effect of an MR image: Robustness and generalizability - p. 465 48. Chapter 17. Scale it up: Large-scale MR data processing - p. 487 49. Chapter 18. Human in the loop: Integration of experts in MR data processing - p. 509 50. Part 6 Clinical application - p. P6 51. Chapter 19. Machine learning applications in neuroradiology - p. 545 52. Chapter 20. AI in cardiovascular MRI - p. 565 53. Chapter 21. Machine learning in body MRI - p. 589 54. Chapter 22. Clinical applications of machine learning in breast MRI - p. 609 55. Chapter 23. Clinical applications of machine learning in musculoskeletal MRI - p. 621 56. Part 7 Reproducibility - p. P7 57. Chapter 24. Let's share: Open-source frameworks and public databases - p. 643 58. Chapter 25. System under test: Challenges for algorithm benchmarking - p. 667 59. Part 8 Conclusion - p. P8 60. Chapter 26. Current solutions, future challenges, and directions - p. 687 61. Index - p. 697
