Color in computer vision: fundamentals and applications

0
958

The book “Color in Computer Vision: Fundamentals and Applications” written by Theo Gevers and published in 2012 by Wiley, is structured in five parts, covering from the fundamentals of color to practical applications in computer vision. It is a work intended both for graduate students and for professionals and academics in the field of computer vision, computer science, and engineering, who possess basic knowledge of image processing, computer vision, and mathematics, including linear algebra, calculus, and probability theory.

In the preface, the authors explain that the project arose from the need to fill a gap in academic texts regarding the integration of color science and computer vision. They highlight the collaboration they have maintained for over a decade at the University of Amsterdam and how this synergy has resulted in a cohesive and consistent work that presents both color theories and essential computational methods for understanding images from a chromatic perspective.

The book begins with a review of the fundamental aspects of color, from trichromatic processing to the similarities between human and computational vision systems. Reflection models that describe the image formation process and its interaction with photometric conditions are discussed, providing a solid foundation for understanding how images are formed and perceived at the computational level.

Once the fundamentals are established, the book explores the extraction of color-invariant information, developing detailed models of the color image formation process and designing mathematical methods to infer quantities of interest. This includes discussions on pixel-based and derivative-based photometric invariance, and provides an overview of the computation of photometric invariance and differential information.

The third part focuses on color constancy, presenting computational methods to estimate illumination and evaluates these methods on large-scale datasets. It addresses how to select and combine different methods to obtain the best possible illumination estimate from a visual scene, employing a statistical approach to quantify unknown data in noisy contexts.

In the fourth part, feature detection and color descriptors are discussed, providing tools for color image processing through an algebraic approach that extends scalar signal processing to vector signals. Computational methods to extract various local image features are introduced here, including circle detectors, curvature estimation, and optical flow.

Finally, the fifth part of the book presents various practical applications, such as image segmentation, object recognition, color naming, and image retrieval. These applications demonstrate how the concepts and methods discussed throughout the book can be implemented to solve real problems in the field of computer vision.

Additionally, the book includes a large amount of supplementary material available online, ranging from software implementations of the presented methods to datasets and slides for academic and research use. The authors invite readers to contribute their own software and datasets to enrich the research community.

Check on Amazon