Presents state-of-the-art algorithms essential for 3D analysis from images
Provides direct algorithm deillegalscriptions without mathematical preliminaries
Includes helpful implementation details for efficient computation and wise memory use
Describes the underlying mathematical theories separately at the end of the volume, and supplies sample codes at an associated website
Dr. Kenichi Kanatani is a Professor Emeritus at Okayama University, Japan. Drs. Yasuyuki Sugaya and Yasushi Kanazawa are Associate Professors in the Department of Computer Science and Engineering at Toyohashi University of Technology, Japan.
This classroom-tested and easy-to-understand textbook/reference describes the state of the art in 3D reconstruction from multiple images, taking into consideration all aspects of programming and implementation. Unlike other computer vision textbooks, this guide takes a unique approach in which the initial focus is on practical application and the procedures necessary to actually build a computer vision system. The theoretical background is then briefly explained afterwards, highlighting how one can quickly and simply obtain the desired result without knowing the derivation of the mathematical detail. Features: reviews the fundamental algorithms underlying computer vision; describes the latest techniques for 3D reconstruction from multiple images; summarizes the mathematical theory behind statistical error analysis for general geometric estimation problems; presents derivations at the end of each chapter, with solutions supplied at the end of the book; provides additional material at an associated website.
Introduction
Part I: Fundamental Algorithms for Computer Vision
Ellipse Fitting
Fundamental Matrix Computation
Triangulation
3D Reconstruction from Two Views
Homography Computation
Planar Triangulation
3D Reconstruction of a Plane
Ellipse Analysis and 3D Computation of Circles
Part II: Multiview 3D Reconstruction
Multiview Triangulation
Bundle Adjustment
Self-calibration of Affine Cameras
Self-calibration of Perspective Cameras
Part III: Mathematical Foundation of Geometric Estimation
Accuracy of Geometric Estimation
Maximum Likelihood and Geometric Estimation
Theoretical Accuracy Limit
Solutions
This classroom-tested and easy-to-understand textbook/reference describes the state of the art in 3D reconstruction from multiple images, taking into consideration all aspects of programming and implementation. Unlike other textbooks on computer vision, this Guide to 3D Vision Computation takes a unique approach in which the initial focus is on practical application and the procedures necessary to actually build a computer vision system. The theoretical background is then briefly explained afterwards, highlighting how one can quickly and simply obtain the desired result without knowing the derivation of the mathematical detail.
Topics and features:
- Reviews the fundamental algorithms underlying computer vision, and their implementation
- Describes the latest techniques for 3D reconstruction from multiple images
- Summarizes the mathematical theory behind statistical error analysis for general geometric estimation problems
- Offers examples of experimental results, enabling the reader to get a feeling of what can be done using each procedure
- Presents derivations and justifications as problems at the end of each chapter, with solutions supplied at the end of the book
- Explains the historical background for each topic in the supplemental notes at the end of each chapter
- Provides additional material at an associated website, include sample code for typical procedures to help readers implement the algorithms described in the book
This accessible work will be of great value to students on introductory computer vision courses. Serving as both as a practical programming guidebook and a useful reference on mathematics for computer vision, it is suitable for practitioners seeking to implement computer vision algorithms as well as for theoreticians wishing to know the underlying mathematical detail.
This classroom-tested and easy-to-understand textbook/reference describes the state of the art in 3D reconstruction from multiple images, taking into consideration all aspects of programming and implementation. Unlike other computer vision textbooks, this guide takes a unique approach in which the initial focus is on practical application and the procedures necessary to actually build a computer vision system. The theoretical background is then briefly explained afterwards, highlighting how one can quickly and simply obtain the desired result without knowing the derivation of the mathematical detail. Features: reviews the fundamental algorithms underlying computer vision; describes the latest techniques for 3D reconstruction from multiple images; summarizes the mathematical theory behind statistical error analysis for general geometric estimation problems; presents derivations at the end of each chapter, with solutions supplied at the end of the book; provides additional material at anassociated website.
Introduction.- Part I: Fundamental Algorithms for Computer Vision.- Ellipse Fitting.- Fundamental Matrix Computation.- Triangulation.- 3D Reconstruction from Two Views.- Homography Computation.- Planar Triangulation.- 3D Reconstruction of a Plane.- Ellipse Analysis and 3D Computation of Circles.- Part II: Multiview 3D Reconstruction.- Multiview Triangulation.- Bundle Adjustment.- Self-calibration of Affine Cameras.- Self-calibration of Perspective Cameras.- Part III: Mathematical Foundation of Geometric Estimation.- Accuracy of Geometric Estimation.- Maximum Likelihood and Geometric Estimation.- Theoretical Accuracy Limit.- Solutions.
Dr. Kenichi Kanatani is a Professor Emeritus at Okayama University, Japan. Drs. Yasuyuki Sugaya and Yasushi Kanazawa are Associate Professors in the Department of Computer Science and Engineering at Toyohashi University of Technology, Japan.
Über den Autor
Dr. Kenichi Kanatani is a Professor Emeritus at Okayama University, Japan. Drs. Yasuyuki Sugaya and Yasushi Kanazawa are Associate Professors in the Department of Computer Science and Engineering at Toyohashi University of Technology, Japan.
Inhaltsverzeichnis
Introduction
Part I: Fundamental Algorithms for Computer Vision
Ellipse Fitting
Fundamental Matrix Computation
Triangulation
3D Reconstruction from Two Views
Homography Computation
Planar Triangulation
3D Reconstruction of a Plane
Ellipse Analysis and 3D Computation of Circles
Part II: Multiview 3D Reconstruction
Multiview Triangulation
Bundle Adjustment
Self-calibration of Affine Cameras
Self-calibration of Perspective Cameras
Part III: Mathematical Foundation of Geometric Estimation
Accuracy of Geometric Estimation
Maximum Likelihood and Geometric Estimation
Theoretical Accuracy Limit
Solutions
Klappentext
This classroom-tested and easy-to-understand textbook/reference describes the state of the art in 3D reconstruction from multiple images, taking into consideration all aspects of programming and implementation. Unlike other computer vision textbooks, this guide takes a unique approach in which the initial focus is on practical application and the procedures necessary to actually build a computer vision system. The theoretical background is then briefly explained afterwards, highlighting how one can quickly and simply obtain the desired result without knowing the derivation of the mathematical detail. Features: reviews the fundamental algorithms underlying computer vision; describes the latest techniques for 3D reconstruction from multiple images; summarizes the mathematical theory behind statistical error analysis for general geometric estimation problems; presents derivations at the end of each chapter, with solutions supplied at the end of the book; provides additional material at anassociated website.
Presents state-of-the-art algorithms essential for 3D analysis from images
Provides direct algorithm deillegalscriptions without mathematical preliminaries
Includes helpful implementation details for efficient computation and wise memory use
Describes the underlying mathematical theories separately at the end of the volume, and supplies sample codes at an associated website
Includes supplementary material: sn.pub/extras
