This book covers the fundamentals in designing and deploying techniques using deep architectures. It is intended to serve as a beginner's guide to engineers or students who want to have a quick start on learning and/or building deep learning systems. This book provides a good theoretical and practical understanding and a complete toolkit of basic information and knowledge required to understand and build convolutional neural networks (CNN) from scratch. The book focuses explicitly on convolutional neural networks, filtering out other material that co-occur in many deep learning books on CNN topics.
| Preface |
|
xi | |
| Acknowledgments |
|
xv | |
| Authors |
|
xvii | |
|
Chapter 1 Introduction to Visual Computing |
|
|
1 | (32) |
|
Image Representation Basics |
|
|
3 | (12) |
|
Transform-Domain Representations |
|
|
6 | (1) |
|
|
|
7 | (3) |
|
Image Gradients and Edges |
|
|
10 | (5) |
|
Going beyond Image Gradients |
|
|
15 | (1) |
|
Line Detection Using the Hough Transform |
|
|
15 | (1) |
|
|
|
16 | (1) |
|
Scale-Invariant Feature Transform |
|
|
17 | (1) |
|
Histogram of Oriented Gradients |
|
|
17 | (8) |
|
Decision-Making in a Hand-Crafted Feature Space |
|
|
19 | (2) |
|
|
|
21 | (2) |
|
Decision-Making with Linear Decision Boundaries |
|
|
23 | (2) |
|
A Case Study with Deformable Part Models |
|
|
25 | (2) |
|
Migration toward Neural Computer Vision |
|
|
27 | (2) |
|
|
|
29 | (1) |
|
|
|
30 | (3) |
|
Chapter 2 Learning as a Regression Problem |
|
|
33 | (32) |
|
|
|
33 | (3) |
|
|
|
36 | (3) |
|
|
|
39 | (2) |
|
Maximum-Likelihood Interpretation |
|
|
41 | (2) |
|
Extension to Nonlinear Models |
|
|
43 | (2) |
|
|
|
45 | (3) |
|
|
|
48 | (1) |
|
|
|
49 | (6) |
|
Geometry of Regularization |
|
|
55 | (2) |
|
|
|
57 | (1) |
|
Stochastic, Batch, and Online Gradient Descent |
|
|
58 | (1) |
|
Alternative Update Rules Using Adaptive Learning Rates |
|
|
59 | (1) |
|
|
|
60 | (2) |
|
|
|
62 | (1) |
|
|
|
63 | (2) |
|
Chapter 3 Artificial Neural Networks |
|
|
65 | (24) |
|
|
|
66 | (8) |
|
Multilayer Neural Networks |
|
|
74 | (5) |
|
The Back-Propagation Algorithm |
|
|
79 | (3) |
|
Improving BP-Based Learning |
|
|
82 | (4) |
|
|
|
82 | (3) |
|
|
|
85 | (1) |
|
|
|
85 | (1) |
|
|
|
86 | (1) |
|
|
|
87 | (2) |
|
Chapter 4 Convolutional Neural Networks |
|
|
89 | (28) |
|
Convolution and Pooling Layer |
|
|
90 | (7) |
|
Convolutional Neural Networks |
|
|
97 | (17) |
|
|
|
114 | (1) |
|
|
|
115 | (2) |
|
Chapter 5 Modern and Novel Usages of CNNs |
|
|
117 | (30) |
|
|
|
118 | (16) |
|
Generality and Transferability |
|
|
121 | (5) |
|
Using Pretrained Networks for Model Compression |
|
|
126 | (4) |
|
Mentee Networks and FitNets |
|
|
130 | (2) |
|
Application Using Pretrained Networks: Image Aesthetics Using CNNs |
|
|
132 | (2) |
|
|
|
134 | (8) |
|
|
|
134 | (3) |
|
Generative Adversarial Networks |
|
|
137 | (5) |
|
|
|
142 | (1) |
|
|
|
143 | (4) |
|
|
|
147 | (12) |
|
|
|
148 | (1) |
|
Quick Start with Yann: Logistic Regression |
|
|
149 | (3) |
|
Multilayer Neural Networks |
|
|
152 | (2) |
|
Convolutional Neural Network |
|
|
154 | (3) |
|
|
|
155 | (2) |
|
|
|
157 | (1) |
|
|
|
157 | (2) |
|
|
|
159 | (4) |
|
|
|
162 | (1) |
| Index |
|
163 | |
Ragav Venkatesan is currently completing his Ph.D. study in Computer Science in the School of Computing, Informatics and Decision Systems Engineering at Arizona State University. He has been a Research Associate with the Visual Representation and Processing Group in ASU, and has worked as a Teaching Assistant for several graduate-level courses in machine learning, pattern recognition, video processing and computer vision. Prior to this, he was a Research Assistant with the Image Processing and Applications Lab in the School of Electrical & Computer Engineering at ASU, where he obtained an M.S. degree in 2012. From 2013 to 2014, Venkatesan was with the Intel Corporation as a computer vision research intern working on technologies for autonomous vehicles. Venkatesan regularly serves as a reviewer for several peer-reviewed journals and conferences in machine learning and computer vision.
Baoxin Li received his Ph.D. in electrical engineering from the University of Maryland, College Park, in 2000. He is currently a Professor and Chair of the Computer Science and Engineering program, and a Graduate Faculty in Electrical Engineering and Computer Engineering programs at Arizona State University, Tempe. From 2000 to 2004, he was a Senior Researcher with SHARP Laboratories of America, Camas, Washington, where he was a technical lead in developing SHARPs trademarked HiMPACT Sports technologies. From 20032004, he was also an Adjunct Professor with the Portland State University, Oregon. He holds eighteen issued U.S. patents and his current research interests include computer vision and pattern recognition, multimedia, social computing, machine learning, and assistive technologies. He won twice the SHARP Laboratories President Award, in 2001 and 2004 respectively. He also won the SHARP Laboratories Inventor of the Year Award in 2002. He was a recipient of the National Science Foundations CAREER Award.
Lisainfo e-raamatute kohta