Skip to main content
Fachgebiete
Automobil + Motoren Bauwesen + Immobilien Business IT + Informatik Elektrotechnik + Elektronik Energie + Nachhaltigkeit Finance + Banking Management + Führung Marketing + Vertrieb Maschinenbau + Werkstoffe Versicherung + Risiko
DE
EN
Bücher
Zeitschriften
Themenseiten
Organisationspsychologie Projektmanagement Marketing Smart Manufacturing
Weitere Formate
Podcasts Webinare Technik Webinare Wirtschaft Kongresse Veranstaltungskalender Awards
Jetzt Einzelzugang starten
Zugang für Unternehmen
Referenzkunden
SLX-Digitalkonferenz: Zukunftswerkstatt 2024

Mehr Umsatz mit Top-Kontakten

Am 7. Mai erfahren Sie, wie Vertriebsteams Leadgenerierung, Leadnurturing und Leadstrategien effizient steuern können.
Erweiterte Suche
Anmelden
nach oben

1995 | Buch

Introduction and Overview Kapitel lesen Erstes Kapitel lesen

Signal Processing for Computer Vision

verfasst von: Gösta H. Granlund, Hans Knutsson

Verlag: Springer US

Enthalten in: Professional Book Archive

Einloggen, um Zugang zu erhalten
insite
SUCHEN

Über dieses Buch

Signal Processing for Computer Vision is a unique and thorough treatment of the signal processing aspects of filters and operators for low-level computer vision.
Computer vision has progressed considerably over recent years. From methods only applicable to simple images, it has developed to deal with increasingly complex scenes, volumes and time sequences. A substantial part of this book deals with the problem of designing models that can be used for several purposes within computer vision. These partial models have some general properties of invariance generation and generality in model generation.
Signal Processing for Computer Vision is the first book to give a unified treatment of representation and filtering of higher order data, such as vectors and tensors in multidimensional space. Included is a systematic organisation for the implementation of complex models in a hierarchical modular structure and novel material on adaptive filtering using tensor data representation.
Signal Processing for Computer Vision is intended for final year undergraduate and graduate students as well as engineers and researchers in the field of computer vision and image processing.

Inhaltsverzeichnis

Frontmatter
Chapter 1. Introduction and Overview
Abstract
Computer vision has progressed considerably over the years. From methods only applicable to simple images, it has developed to deal with increasingly complex scenes, volumes and time sequences.
Gösta H. Granlund, Hans Knutsson
Chapter 2. Biological Vision
Abstract
The most, some would say the only, successful vision systems are the biological ones, ranging in complexity from those found in lower animals to the very flexible and highly integrated system for human vision. Although it has not, in general, proved feasible to employ biological functions for technical systems, the situation may be quite different in the field of information processing. The computational structures which have been developed in our current computer technology are primarily designed to handle information in strings. This is very different from spatial and cognitive information, and it has been found that the current organization of computers does not support an efficient representation and computation for spatial data. For this reason it is of great interest to look at the organization of biological visual systems. The study of such systems may suggest more efficient information representations, computation structures and suitable primitives for representation of spatial information.
Gösta H. Granlund, Hans Knutsson
Chapter 3. Low Level Operations
Abstract
Image computers store and process large chunks of data representing image information in various forms. This is usually done in the form of arrays, which are contiguous areas of memory with data elements laid out one after the other. The storage layout of arrays simplifies data access when all items are elements in regular arrays. Then the program merely needs to take the position of the element and multiply the y-coordinate by the row length, and to add the x-coordinate and the start address of the array. For images or data arrays with a fixed size, this is a convenient representation.
Gösta H. Granlund, Hans Knutsson
Chapter 4. Fourier Transforms
Abstract
This chapter presents an overview of definitions and theorems for Fourier transforms in the context of computer vision. The presentation is made without any claims on being either thorough or mathematically strict. Some results are stated without proofs and the transforms of various functions are assumed to exist without further ado. For a more rigorous treatment of these issues, see, for example, [23]. Furthermore, it is assumed that the reader is familiar with the one-dimensional Fourier transform, both the continuous version and the discrete Fourier transform, as well as some mathematical constructions such as generalized functions or distributions.
Gösta H. Granlund, Hans Knutsson
Chapter 5. Kernel Optimization
Abstract
In this chapter it is shown how convolution kernels can be implemented in practical situations. The material is based on [65]. The presentation is restricted to the case where the desired ideal filter is given in the Fourier domain. An optimal kernel is the set of coefficients that minimizes some distance measure with respect to the ideal filter. A family of distance measures suitable for multidimensional image signals is given. The dependence of attainable error levels on kernel size is demonstrated and convolution results on test images discussed.
Gösta H. Granlund, Hans Knutsson
Chapter 6. Orientation and Velocity
Abstract
This chapter presents a unified approach and a theory for estimation of local orientation and velocity in multi-dimensional spatial or spatio-temporal signals. The presentation is intended to serve as a basis for the design of efficient multidimensional signal analysis algorithms.
Gösta H. Granlund, Hans Knutsson
Chapter 7. Local Phase Estimation
Abstract
Most people are familiar with the global Fourier phase. The shift theorem, describing how the Fourier phase is affected by moving the signal, is common knowledge. The phase in signal representations based on local operations, e.g. lognormal filters, is on the other hand not so well known.
Gösta H. Granlund, Hans Knutsson
Chapter 8. Local Frequency
Abstract
This chapter describes a technique for estimation of local signal frequency and bandwidth. Local frequency is an important concept useful for local structure analysis as well as for determining the appropriate range of scales for subsequent processing. The method is based on combining local estimates of instantaneous frequency over a large number of scales. The filters used are a set of lognormal quadrature filters. The bandwidth is used to produce a measure of certainty for the estimated frequency. The algorithm is applicable to multidimensional data and examples of the performance of the method are demonstrated for one-dimensional and two-dimensional signals.
Gösta H. Granlund, Hans Knutsson
Chapter 9. Representation and Averaging
Abstract
Information representation is an important and complex issue in all multi-level signal processing systems [42]. In this chapter we discuss specific properties that are claimed to be of major importance for representation of information in visual processing systems.
Gösta H. Granlund, Hans Knutsson
Chapter 10. Adaptive Filtering
Abstract
This chapter presents a computationally efficient technique for adaptive filtering of N-dimensional signals. The approach is based on the local signal description given by the orientation tensor discussed in chapter 6. The adaptive filter is synthesized as a tensor-controlled weighted summation of shift-invariant filters.
Gösta H. Granlund, Hans Knutsson
Chapter 11. Vector and Tensor Field Filtering
Abstract
In the preceding chapters, we have been concerned with operations acting on multi-dimensional scalar images. Recall, for example, the orientation algorithm estimating the local orientation in Chapter 6 and the algorithm estimating local frequency in Chapter 8. In this chapter, we describe some algorithms acting on vectors and higher order tensors.
Gösta H. Granlund, Hans Knutsson
Chapter 12. Classification and Response Generation
Abstract
The last stage of a computer vision system generally produces some action in the external world. It may be to control the motors of an industrial robot. It may be to point out defective units on the conveyor belt of a quality control system. Or it may be to point out suspected cancerous regions in an X-ray image. Sometimes the input to the system is used to improve a robot’s knowledge and model of the world without any immediate external response being generated. Still, the conceptual purpose of a computer vision system is to produce responses, or at least to represent visual data in such a form that a correct decision is made.
Gösta H. Granlund, Hans Knutsson
Chapter 13. Texture Analysis
Abstract
Texture is the term used to characterize the surface of a given object and it is unquestionably one of the main features used in computer vision and pattern recognition. Computer vision researchers have used measures of texture to segment scenes and discriminate between different objects. Psychophysicists have studied texture for knowledge about human low-level visual information processing. It has been assumed that a solution to the texture analysis problem will greatly advance the computer vision and pattern recognition field. But, in spite of its importance, texture lacks a precise definition.
Gösta H. Granlund, Hans Knutsson
Backmatter
Metadaten
Titel
Signal Processing for Computer Vision
verfasst von
Gösta H. Granlund
Hans Knutsson
Copyright-Jahr
1995
Verlag
Springer US
Electronic ISBN
978-1-4757-2377-9
Print ISBN
978-1-4419-5151-9
DOI
https://doi.org/10.1007/978-1-4757-2377-9