Computer Vision Metrics: Survey, Taxonomy, and Analysis

by Scott Krig

This new textbook edition provides a comprehensive history and state-of-the-art survey for fundamental computer vision methods. Expanded and updated, this book features over 300 new references, totaling over 800 in all, as well as learning assignments at the end of each chapter to help students and researchers dig deeper into key topics. This survey covers everything from imaging devices, computational imaging, interest point detectors, local feature descriptors, regional and global feature metrics, feature learning architectures, deep learning, neuroscience of vision, neural networks, and detailed example architectures to illustrate computer vision hardware and software optimization methods. To complement the survey, the book includes useful analysis to provide intuition into the goals of various methods, why they work, and how they may be optimized. This is not a how-to book with source code examples, but rather a survey and taxonomy intended as a reference tool for researchers and engineers, complimenting the many fine hand-on resources and open source projects such as OpenCV and other imaging and deep learning tools.

740 pages, Hardcover

First published May 27, 2014