Local Image Descriptor: Modern Approaches
Published on 4. January 2016
XII, 99 pages
978-3-662-49173-7 (ISBN)
Description
This book covers a wide range of local image descriptors, from the classical ones to the state of the art, as well as the burgeoning research topics on this area. The goal of this effort is to let readers know what are the most popular and useful methods in the current, what are the advantages and the disadvantages of these methods, which kind of methods is best suitable for their problems or applications, and what is the future of this area. What is more, hands-on exemplars supplied in this book will be of great interest to Computer Vision engineers and practitioners, as well as those want to begin their research in this area. Overall, this book is suitable for graduates, researchers and engineers in the related areas both as a learning text and as a reference book.
More details
Series
Edition
1st ed. 2015
Language
English
Place of publication
Berlin
Germany
Publishing group
Springer Berlin
Target group
Primary & secondary/elementary & high school
Product notice
Reflowable
Illustrations
XII, 99 p. 33 illus., 7 illus. in color.
File size
3,38 MB
ISBN-13
978-3-662-49173-7 (9783662491737)
DOI
10.1007/978-3-662-49173-7
Schweitzer Classification
Other editions
Additional editions
Bin Fan | Zhenhua Wang | Fuchao Wu
Local Image Descriptor: Modern Approaches
Book
02/2016
Springer
€53.49
Shipment within 7-9 days
Persons
Dr. Bin Fan is an Associate Professor in the National Laboratory of Pattern Recognition, Institute of Automation, Chinese Academy of Sciences where he got his PhD degree in 2011 and acted as an Assistant Professor from 2011 to 2014. He has been a visiting professor at the Computer Vision Laboratory, Ecole Polytechnique Federale de Lausanne (EPFL), Switzerland, during 2014.5 to 2014.6 and 2015.3 to 2016.3. His research interests include computer vision and pattern recognition. Particularly, he is focusing on designing and matching of local features and their visual applications. He has published over 20 papers in prestigious international journals and conferences such as IEEE TPAMI, IEEE TIP, PR, CVPR, ICCV, ECCV, AAAI. He serves as an Associate Editor of Neurocomputing, Area Chair of WACV'16 and regular reviewer for top-ranking journals as well as on program committee member for major vision conferences. In CVPR'15, he was awarded the Outstanding Reviewer Awards and is currently amember of IEEE. He gave a tutorial about local invariant descriptors in the Vision and Learning Seminar (VALSE) 2014.
Dr. Zhenhua Wang is a research fellow in the Rapid-Rich Object Search (ROSE) Lab, School of EEE, Nanyang Technological University, Singapore since 2014.8. He received his BS degree in software engineering from Sichuan University in 2008, and the PhD degree in pattern analysis and machine intelligence from the National Laboratory of Pattern Recognition, Institute of Automation, Chinese Academy of Sciences. His research interests are in the fields of Computer Vision related topics, including feature detection, feature description, 3D reconstruction.
Prof. Fuchao Wu is a Professor in the National Laboratory of Pattern Recognition, Institute of Automation, Chinese Academy of Sciences. Previously, he acted as a Lecturer and then as an Associate Professor in Anqing Teacher's College from 1984 to 1994. From 1995 to 2000, he acted as a Professor in Anhui University. His research interests are now in computer vision, which includes multi-view geometry, 3D reconstruction, active vision, and image based modeling. He has published two books and over 100 papers in prestigious international journals and conferences such as IJCV, IEEE TPAMI, IEEE TIP, CVIU, PR, CVPR, ICCV, ECCV etc. He has received several honors and awards, including the Second Best Award of Natural Science of Anhui province in 2000.
Dr. Zhenhua Wang is a research fellow in the Rapid-Rich Object Search (ROSE) Lab, School of EEE, Nanyang Technological University, Singapore since 2014.8. He received his BS degree in software engineering from Sichuan University in 2008, and the PhD degree in pattern analysis and machine intelligence from the National Laboratory of Pattern Recognition, Institute of Automation, Chinese Academy of Sciences. His research interests are in the fields of Computer Vision related topics, including feature detection, feature description, 3D reconstruction.
Prof. Fuchao Wu is a Professor in the National Laboratory of Pattern Recognition, Institute of Automation, Chinese Academy of Sciences. Previously, he acted as a Lecturer and then as an Associate Professor in Anqing Teacher's College from 1984 to 1994. From 1995 to 2000, he acted as a Professor in Anhui University. His research interests are now in computer vision, which includes multi-view geometry, 3D reconstruction, active vision, and image based modeling. He has published two books and over 100 papers in prestigious international journals and conferences such as IJCV, IEEE TPAMI, IEEE TIP, CVIU, PR, CVPR, ICCV, ECCV etc. He has received several honors and awards, including the Second Best Award of Natural Science of Anhui province in 2000.
Content
- Intro
- Foreword 1
- Foreword 2
- Preface
- Contents
- 1 Introduction
- References
- 2 Classical Local Descriptors
- 2.1 Scale-Invariant Feature Transform (SIFT)
- 2.1.1 Scale Space Representation in SIFT
- 2.1.2 Keypoint Detection
- 2.1.3 Feature Description
- 2.2 Speeded Up Robust Feature (SURF)
- 2.2.1 Integral Image
- 2.2.2 Scale Space Representation in SURF
- 2.2.3 Scale-Invariant Interest Point Detection
- 2.2.4 Orientation Assignment and Descriptor Construction
- 2.3 Local Binary Pattern and Its Variants
- References
- 3 Intensity Order-Based Local Descriptors
- 3.1 Ordinal and Spatial Intensity Distribution Descriptor (OSID)
- 3.2 Intensity Order-Based Pooling for Feature Description
- 3.2.1 An Analysis of the Geometric-Based Spatial Pooling
- 3.2.2 Intensity Order-Based Patch Division
- 3.2.3 Construction of MROGH and MRRID Descriptors
- 3.3 Local Intensity Order Pattern for Feature Description
- 3.3.1 Construction of the LIOP Descriptor
- 3.4 Intensity Order-Based Binary Descriptor
- 3.4.1 Subregions Generation
- 3.4.2 Regional Invariants and Pairwise Comparisons
- 3.4.3 Learning Good Binary Descriptor
- 3.4.4 Using Multiple Support Regions
- 3.4.5 Cascade Filtering for Speeding up Matching
- References
- 4 Burgeoning Methods: Binary Descriptors
- 4.1 BRIEF: Binary Robust Independent Elementary Features
- 4.2 ORB: Oriented FAST and Rotated BRIEF
- 4.2.1 Scale Invariant FAST Detector
- 4.2.2 Orientation Computation by Intensity Centriod
- 4.2.3 Learning Good Binary Features
- 4.3 BRISK: Binary Robust and Invariant Scalable Keypoints
- 4.3.1 Keypoint Detection
- 4.3.2 Orientation Assignment and Keypoint Description
- 4.4 FREAK: Fast Retina Keypoint
- 4.4.1 Descriptor Construction
- 4.4.2 Saccadic Matching with FREAK
- 4.5 FRIF: Fast Robust Invariant Feature
- 4.5.1 FALoG Detector
- 4.5.2 Mixed Binary Descriptor
- 4.6 Learning Binary Descriptors by Supervised Information
- 4.6.1 From Raw Image Patch
- 4.6.2 From an Intermediate Representation
- References
- 5 Visual Applications
- 5.1 Structure from Motion and 3D Reconstruction
- 5.2 Object Recognition
- 5.3 Content-Based Image Retrieval
- 5.4 Simultaneous Localization and Mapping (SLAM)
- References
- 6 Resources and Future Work
- 6.1 Dataset and Evaluation Protocol
- 6.1.1 Benchmarks for Image Matching
- 6.1.2 Benchmarks for Object Recognition
- 6.1.3 Benchmarks for Image Retrieval
- 6.2 Conclusion Remarks and Future Work
- References
