Foundations of Computer Vision

Computational Geometry, Visual Image Structures and Object Shape Detection

verfasst von: James F. Peters

Verlag: Springer International Publishing

Buchreihe : Intelligent Systems Reference Library

Enthalten in: Springer Professional "Wirtschaft+Technik" , Springer Professional "Technik" , Springer Professional "Wirtschaft"

Einloggen, um Zugang zu erhalten

Über dieses Buch

This book introduces the fundamentals of computer vision (CV), with a focus on extracting useful information from digital images and videos. Including a wealth of methods used in detecting and classifying image objects and their shapes, it is the first book to apply a trio of tools (computational geometry, topology and algorithms) in solving CV problems, shape tracking in image object recognition and detecting the repetition of shapes in single images and video frames. Computational geometry provides a visualization of topological structures such as neighborhoods of points embedded in images, while image topology supplies us with structures useful in the analysis and classiﬁcation of image regions. Algorithms provide a practical, step-by-step means of viewing image structures.

The implementations of CV methods in Matlab and Mathematica, classiﬁcation of chapter problems with the symbols (easily solved) and (challenging) and its extensive glossary of key words, examples and connections with the fabric of CV make the book an invaluable resource for advanced undergraduate and ﬁrst year graduate students in Engineering, Computer Science or Applied Mathematics.

It offers insights into the design of CV experiments, inclusion of image processing methods in CV projects, as well as the reconstruction and interpretation of recorded natural scenes.

Inhaltsverzeichnis

Frontmatter

Chapter 1. Basics Leading to Machine Vision

Abstract

The principal aim of computer vision is to reconstruct and interpret natural scenes based on the content of images captured by digital cameras [190]. A natural scene is that part of visual field that is captured either by human visual perception or by optical sensor arrays.

James F. Peters

Chapter 2. Working with Pixels

Abstract

A pixel (aka picture element) is an element at position (r, c) (row, column) in a digital image I . A pixel represents the smallest constituent element in a digital image. Typically, each pixel in a raster image is represented by a tiny square called a raster image tile. Raster image technology has its origins in the raster scan of cathode ray tube (CRT) displays in which images are rendered line-by-line by magnetically steering a focused electron beam. Usually, computer monitors have bitmapped displays in which each screen pixel corresponds to its bit depth, i.e., number of pixels used to render pixel colour channels.

James F. Peters

Chapter 3. Visualising Pixel Intensity Distributions

Abstract

This chapter introduces various ways to visualize pixel intensity distributions. Also included here are pointers to sources of generating points useful in image tessellations and triangulations. In other words, image structure visualizations carries with it tacit insights about image geometry.

James F. Peters

Chapter 4. Linear Filtering

Abstract

This chapter introduces linear spatial filters. A linear filter is a time-invariant device (function, or method) that operates on a signal to modify the signal in some fashion.

James F. Peters

Chapter 5. Edges, Lines, Corners, Gaussian Kernel and Voronoï Meshes

Abstract

This chapter focuses on the detection of edges, lines and corners in digital images. This chapter also introduces a number of non-linear filtering methods. A method is a non-linear method, provided the output of the method is not directly proportional to the input. For example, a method whose input is a real-valued variable x and whose output is \(x^{\alpha }, \alpha > 0\) (power of x) is non-linear.

James F. Peters

Chapter 6. Delaunay Mesh Segmentation

Abstract

This chapter introduces segmentation of digital images using Delaunay meshes. An image is segmented by separating the image into almost disjoint regions. The interiors of image segments do not overlap. Each segment contains points that belong only to the segment.

James F. Peters

Chapter 7. Video Processing. An Introduction to Real-Time and Offline Video Analysis

Abstract

This chapter introduces video processing with the focus on tracking changes in video frame images. Video frame changes can be detected in the changing shapes, locations and distribution of the polygons (regions) in Voronoï tilings of the frames.

James F. Peters

Chapter 8. Lowe Keypoints, Maximal Nucleus Clusters, Contours and Shapes

Abstract

This chapter carries forward the use of Voronoï meshes superimposed on digital images as a means revealing image geometry and the shapes that result from contour lines surrounding maximal nucleus clusters (MNCs) in a mesh.

James F. Peters

Chapter 9. Postscript. Where Do Shapes Fit into the Computer Vision Landscape?

Abstract

Shapes are elusive creatures that drift in and out of natural scenes that we sometimes perceive, store in memory and record with digital cameras.

James F. Peters

Backmatter

Titel: Foundations of Computer Vision
verfasst von: James F. Peters
Verlag: Springer International Publishing
Electronic ISBN: 978-3-319-52483-2
Print ISBN: 978-3-319-52481-8
DOI: https://doi.org/10.1007/978-3-319-52483-2