Free application: Deep Learning and Medical Image Processing

EEWORLD社区

Free application: Deep Learning and Medical Image Processing [Copy link]

This is an introductory book on " how to use deep learning methods to solve medical image processing problems ". This book first introduces the basic knowledge of medical images, including medical image data, data annotation, medical digital image processing, and medical image classification; secondly, it introduces the solution of common machine vision tasks in medical image processing (semantic segmentation, key point detection, and medical image registration), supplemented by practical cases, to help readers deeply understand the relevant technical principles and consolidate the knowledge they have learned; then it introduces the relevant content of model optimization and transfer learning to help readers broaden their thinking and improve their ability to adopt different solutions for specific needs.
This book is suitable for engineers and researchers in medical and engineering majors and those engaged in medical image processing. It can also be used as a reference for senior undergraduates and graduate students in intelligent medicine related majors. Before reading this book, readers need to understand basic deep learning knowledge and have a certain foundation in Python programming .

Editor's recommendation:

1. A wonderful combination of medicine and engineering, detailing how to use deep learning methods to solve medical image processing problems.
2. Written by the team of the National Clinical Research Center for Neurological Diseases, Beijing Tiantan Hospital, Capital Medical University, and recommended by the dean Wang Yongjun.
3. Rich and practical content. It covers the relevant models and key technologies in various application directions in this field, including Transformer, BERT, GPT, etc.
4. Equipped with example code, it is runnable, practical, and the language is easy to understand. It is suitable as a reference book for getting started with natural language processing, and can also be used as a practical guide for professionals working in this field.
5. Detailed explanation of real practical cases in medical projects, easy to learn and use. It brings together important ideas and research results from many papers, sorts out the development context of natural language processing technology, and better understands the development trend of this field.

how to apply

(1) Check the product information below to learn more about the book content and chapter settings;

(2) >> Click "I want to apply " and fill in the reason for application, reading sharing plan, etc. carefully to have a chance to get the book for free. There is 1 quota for the evaluation voucher in this activity, first come first served.

Event Schedule

Application period: February 20 to March 10
Selection announcement: All shortlisted candidates will be announced before March 13
Reading period: March 20 to April 30
Awarding time: within two weeks after the event ends

Product information

About the Author

Liang Longkai, Master of Computer Science and Technology from Harbin University of Science and Technology, Senior Algorithm Engineer of Brain Science and Brain-like Research Department of Changping National Laboratory, Distinguished Senior Algorithm Engineer of Artificial Intelligence Research and Development Project Group of National Clinical Medical Research Center for Neurological Diseases. Fu He, Master of Robotics Technology from Beihang University, Distinguished Senior Algorithm Engineer of Artificial Intelligence Research and Development Project Group of National Clinical Medical Research Center for Neurological Diseases, Technical Advisor of the General Hospital of the People's Liberation Army, Algorithm Director of Artificial Intelligence Department of Biomind. Chen Fengwei, Master of Dalian University of Technology, Algorithm Engineer of Brain Science and Brain-like Research Department of Changping National Laboratory, Distinguished Algorithm Engineer of Artificial Intelligence Research and Development Project Group of National Clinical Medical Research Center for Neurological Diseases. Liu Yaou, Member of the Party Committee of Beijing Tiantan Hospital Affiliated to Capital Medical University, Discipline Leader of Radiology Department (National Key Clinical Specialty), Chief Physician, Professor, and Doctoral Supervisor. Xiong Yunyun, Graduated from Fudan University's Seven-Year Clinical Medicine Program (Undergraduate and Master), Doctor and Postdoctoral Fellow of the Chinese University of Hong Kong, Visiting Researcher at Harvard University. Currently Chief Physician, Associate Professor, Master Supervisor, and Deputy Director of Vascular Neurology Department of Neurology Center, Beijing Tiantan Hospital Affiliated to Capital Medical University.

Book Catalog

Chapter 1 Application of Artificial Intelligence in Medicine 1
1.1 Overview of Artificial Intelligence 1
1.2 Application of Artificial Intelligence in Medicine 2
1.3 Application of Artificial Intelligence in Medical Imaging 5
1.4 Summary 6
1.5 References 7


Chapter 2 Medical Image Data 11
2.1 Common Medical Image Data 11
2.1.1 X-ray Imaging 11
2.1.2 X-ray Computed Tomography 12
2.1.3 Magnetic Resonance Imaging 13
2.1.4 Ultrasound Imaging 16
2.1.5 Electrocardiogram 17
2.2 Common Image Formats 18
2.2.1 DICOM 18
2.2.2 Analyze 23
2.2.3 Nifti 24
2.2.4 Minc 25
2.3 Summary 25
2.4 References 25

Chapter 3 Data Annotation 27
3.1 Interface Introduction 27
3.2 Start Annotation 31
3.3 Summary 41
3.4 References 41

Chapter 4 Medical Digital Image Processing 43
4.1 Data Preprocessing 44
4.1.1 Interpolation 44
4.1.2 Resampling 46
4.1.3 Analysis and Equalization of Signal Intensity Histogram 48
4.1.4 Data Normalization 50
4.1.5 Connected Domain Analysis 51
4.1.6 Morphological Methods 52
4.2 Data Enhancement 55
4.2.1 Common Data Enhancement Methods 55
4.2.2 Elastic Deformation 56
4.2.3 Online Data Enhancement Based on TensorFlow 57
4.3 Summary 59
4.4 References 59

Chapter 5 Medical Image Classification 61
5.1 Loss Function 61
5.1.1 Cross Entropy Loss 62
5.1.2 Focal Loss 62
5.1.3 KL Divergence 63
5.2 Evaluation Indicators 64
5.2.1 Confusion Matrix 64
5.2.2 Common evaluation indicators 64
5.2.3 Common evaluation indicators for diagnostic experiments 67
5.2.4 Evaluation indicators for measuring model performance 67
5.3 Classic model 68
5.3.1 Cross-layer connection 69
5.3.2 Network width 71
5.3.3 Attention mechanism 72
5.4 Practice: Classification and detection of cerebral hemorrhage based on intracranial CT images 73
5.4.1 Dataset preprocessing 74
5.4.2 Model training 78
5.4.3 Model testing 84
5.4.4 Practical summary of classification and detection of cerebral hemorrhage based on intracranial CT images 86
5.5 Summary 86
5.6 References 86

Chapter 6 Semantic Segmentation 89
6.1 Loss function 89
6.1.1 Dice loss 90
6.1.2 Tversky loss 90
6.1.3 Boundary loss 91
6.1.4 Mixed loss 91
6.2 6.2.1
IoU 92
6.2.2 Dice coefficient 93
6.2.3 Hausdorff-95 93
6.3 Other statistical methods 94
6.3.1 Patient-level 94
6.3.2 Data-level 94
6.4 Classic segmentation model 95
6.4.1 UNet network 95
6.4.2 UNet deformation 97
6.4.3 Other segmentation networks 99
6.5 Practice: Brain tumor segmentation based on MRI images 100
6.5.1 Data preprocessing 100
6.5.2 Model building 104
6.5.3 Training model 108
6.5.4 Model testing 113
6.6 Summary 114
6.7 References 115

Chapter 7 Key point detection 117
7.1 Concept and significance 117
7.2 Common key point detection models 118
7.3 7.4
Summary 130
7.5 References 130

Chapter 8 Medical Image Registration 131
8.1 Basics 131
8.1.1 Feature Space 132
8.1.2 Search Space 132
8.1.3 Similarity Metrics 136
8.1.4 Search Strategy 138
8.1.5 Quality Assessment 139
8.2 Deep Learning Image Registration Methods 140
8.2.1 Supervised Learning Image Registration 141
8.2.2 Unsupervised Learning Image Registration 142
8.3 Practice: Deep Learning Image Registration Model VoxelMorph 142
8.3.1 Data Reading 143
8.3.2 Network Structure 144
8.3.3 Training and Testing 149
8.3.4 Practical Summary 151
8.4 Summary 151
8.5 References 152

Chapter 9 Model Optimization 153
9.1 Model Pruning 153
9.1.1 Concept of Sparsity 154
9.1.2 Pruning Strategy 154
9.1.3 Sensitivity Analysis 156
9.2 Model Quantization 157
9.3 TensorRT 158
9.3.1 Basic Introduction 158
9.3.2 Application Scenarios 158
9.3.3 Basic Principles 159
9.4 Practical: Quantization of Intracranial Hemorrhage CT Image Classification Model 160
9.5 Summary 163
9.6 References 163

Chapter 10 Transfer Learning 165
10.1 Transfer Learning 165
10.2 Lifelong Learning 166
10.3 Practical: Optimization Method for Intracranial Hemorrhage Detection with Imbalanced Data 167
10.3.1 Experiments in Transfer Learning 167
10.3.2 Experiments in Lifelong Learning 177
10.4 Summary 184
10.5 References 184

For more event content, click here to enter the event page

ew123

Looking at the introduction, it seems to be quite up-to-date. However, this book is just for beginners, after all, medical images are still very strict.

吾妻思萌

Interdisciplinary medical disciplines have been very popular in recent years and are worth studying.

秦天qintian0303

It's a bit complicated. It requires a lot of medical data to support it. It also requires both hardware and software.

lugl4313820

I am actually a suitable reader for this book.

1. I have a medical background.

2. 5 years of Python development experience.

3. Full-stack development engineer.

4. Have a colleague with CT scan.

But I am not very familiar with AI.