This book offers an overview of traditional big visual data analysis approaches and provides state-of-the-art solutions for several scene comprehension problems, indoor/outdoor classification, outdoor scene classification, and outdoor scene layout estimation. It is illustrated with numerous natural and synthetic color images, and extensive statistical analysis is provided to help readers visualize big visual data distribution and the associated problems. Although there has been some research on big visual data analysis, little work has been published on big image data distribution analysis using the modern statistical approach described in this book. By presenting a complete methodology on big visual data analysis with three illustrative scene comprehension problems, it provides a generic framework that can be applied to other big visual data analysis tasks.
|
|
|
1 | (6) |
|
|
|
3 | (4) |
|
2 Scene Understanding Datasets |
|
|
7 | (16) |
|
2.1 Small-Scale Scene Understanding Datasets |
|
|
7 | (3) |
|
|
|
7 | (1) |
|
|
|
8 | (1) |
|
|
|
9 | (1) |
|
|
|
9 | (1) |
|
2.2 Large-Scale Scene Understanding Datasets |
|
|
10 | (13) |
|
2.2.1 80 Million Tiny Image Dataset |
|
|
11 | (1) |
|
|
|
12 | (1) |
|
|
|
13 | (1) |
|
|
|
13 | (1) |
|
2.2.5 Scene Understanding (SUN) Dataset |
|
|
14 | (5) |
|
|
|
19 | (1) |
|
|
|
20 | (3) |
|
3 Indoor/Outdoor Classification with Multiple Experts |
|
|
23 | (42) |
|
|
|
23 | (2) |
|
3.2 Individual Indoor/Outdoor Experts |
|
|
25 | (21) |
|
3.2.1 Experts from Existing Work |
|
|
25 | (16) |
|
|
|
41 | (5) |
|
3.3 Data Grouping Using Experts' Decisions |
|
|
46 | (2) |
|
3.4 Diversity Gain of Experts Via Decisions Stacking |
|
|
48 | (5) |
|
3.4.1 Diversity Gain of Two Experts |
|
|
49 | (3) |
|
3.4.2 Construction of Multi-Expert Systems |
|
|
52 | (1) |
|
3.5 Expert Decision Fusion Systems |
|
|
53 | (1) |
|
3.6 Performance Evaluation |
|
|
54 | (7) |
|
3.6.1 Performance of Individual Expert |
|
|
54 | (3) |
|
3.6.2 Subspace Classification Performance |
|
|
57 | (1) |
|
|
|
58 | (2) |
|
|
|
60 | (1) |
|
|
|
61 | (4) |
|
|
|
61 | (4) |
|
4 Outdoor Scene Classification Using Labeled Segments |
|
|
65 | (28) |
|
|
|
65 | (2) |
|
4.2 Review of Previous Works |
|
|
67 | (12) |
|
|
|
67 | (2) |
|
|
|
69 | (3) |
|
4.2.3 High-Level Features |
|
|
72 | (1) |
|
|
|
73 | (2) |
|
4.2.5 Scene Parsing and Semantic Segmentation |
|
|
75 | (4) |
|
4.3 Proposed Coarse Semantic Segmentation (CSS) |
|
|
79 | (6) |
|
4.3.1 Limitations of Traditional Learning Units |
|
|
79 | (1) |
|
4.3.2 Coarse Segmentation |
|
|
80 | (2) |
|
4.3.3 Segmental Semantic Labeling |
|
|
82 | (3) |
|
4.4 Scene Classification Using CSS |
|
|
85 | (1) |
|
|
|
86 | (3) |
|
|
|
86 | (1) |
|
|
|
86 | (1) |
|
4.5.3 Scene Classification Results |
|
|
87 | (2) |
|
|
|
89 | (4) |
|
|
|
90 | (3) |
|
5 Global-Attributes Assisted Outdoor Scene Geometric Labeling |
|
|
93 | (28) |
|
|
|
93 | (1) |
|
5.2 Review of Previous Works |
|
|
94 | (5) |
|
5.2.1 Geometric Context from a Single Image |
|
|
94 | (1) |
|
5.2.2 Blocks World Revisited |
|
|
95 | (2) |
|
5.2.3 Single-View 3D Scene Parsing by Attributed Grammar |
|
|
97 | (1) |
|
5.2.4 Inferring 3D Layout of Buildings from a Single Image |
|
|
98 | (1) |
|
|
|
99 | (10) |
|
|
|
99 | (1) |
|
5.3.2 Initial Pixel Labeling (IPL) |
|
|
100 | (1) |
|
5.3.3 Global Attributes Extraction (GAE) |
|
|
101 | (7) |
|
5.3.4 Layout Reasoning and Label Refinement (LR2) |
|
|
108 | (1) |
|
|
|
109 | (9) |
|
|
|
118 | (3) |
|
|
|
119 | (2) |
|
6 Conclusion and Future Work |
|
|
121 | |
Chen Chen received his B.S. degree in Electrical Engineering from Beijing University of Posts and Telecommunications (BUPT) in 2010. He received his M.S. degree in Electrical Engineering from University of Southern California (USC) in 2012. At the same year, he joined the Media Communication Lab led by Professor Kuo in University of Southern California (USC), where he is pursuing her Ph.D degree in Electrical Engineering and serving as a research assistant. His research interests include image classification, image tagging and image/video processing.
Yu-Zhuo Ren received her B.S. degree in Hebei University of Technology (HUT), China, in 2011 and the M.S. degree in Electrical Engineering from University of Southern California (USC) in 2013. She is now working as a research assistant in the Media Communication Lab led by Professor Kuo. Her research interests include image understanding related problems, in the field of computer vision and machine learning.
C.-C. Jay Kuo Dr. C.-C. Jay Kuo received the B.S. degree from the National Taiwan University, Taipei, in 1980 and the M.S. and Ph.D. degrees from the Massachusetts Institute of Technology, Cambridge, in 1985 and 1987, respectively, all in Electrical Engineering. From October 1987 to December 1988, he was Computational and Applied Mathematics Research Assistant Professor in the Department of Mathematics at the University of California, Los Angeles. Since January 1989, he has been with the University of Southern California (USC). He is presently Director of the Multimedia Communication Lab. and Professor of Electrical Engineering and Computer Science at the USC. His research interests are in the areas of multimedia data compression, communication and networking, multimedia content analysis and modeling, and information forensics and security. Dr. Kuo has guided 119 students to their Ph.D. degrees and supervised 23 postdoctoral research fellows. Currently, his research group at the USC has around 30Ph.D. students, which is one of the largest academic research groups in multimedia technologies. He is coauthor of about 220 journal papers, 850 conference papers and 12 books. He delivered over 550 invited lectures in conferences, research institutes, universities and companies.
Lisainfo e-raamatute kohta