Computational Imaging and Vision – serie

Scale is a concept the antiquity of which can hardly be traced. Certainly the familiar phenomena that accompany sc ale changes in optical patterns are mentioned in the earliest written records. The most obvious topological changes such as the creation or annihilation of details have been a topic to philosophers, artists and later scientists. This appears to of fascination be the case for all cultures from which extensive written records exist. For th instance, chinese 17 c artist manuals remark that "distant faces have no eyes" . The merging of details is also obvious to many authors, e. g. , Lucretius mentions the fact that distant islands look like a single one. The one topo­ logical event that is (to the best of my knowledge) mentioned only late (by th John Ruskin in his "Elements of drawing" of the mid 19 c) is the splitting of a blob on blurring. The change of images on a gradual increase of resolu­ tion has been a recurring theme in the arts (e. g. , the poetic description of the distant armada in Calderon's The Constant Prince) and this "mystery" (as Ruskin calls it) is constantly exploited by painters.

This volume contains the papers presented at the 14th International Conference on Information Processing in Medical Imaging. IPMI meetings have a a strong emphasis on the clinical relevance and validation of medical imaging. This book covers the whole spectrum: acquisition, tomographic reconstruction, registration, segmentation, knowledge-based analysis, display and image quality as well as several important applications. Several papers present significant advances in topics already discussed at previous meetings while others deal with new topics and methodology, opening new horizons in medical imaging. In addition to the 28 full-length papers, 30 short communications are included to sample the most current work in progress. Audience: An up-to-date and complete overview of ongoing research in medical imaging, beneficial to all physicists, computer scientists and physicians who wish to remain informed on state-of-the-art methodology in medical imaging.

This book contains twenty-one selected papers based on communications presented at the "Third International Meeting on Fully Three-Dimensional Image Reconstruction in Radiology and Nuclear Medicine", held July 4-6, 1995 at Domaine d'Aix- Marlioz, Aix-les-Bains, France. 3D tomographic imaging systems based on ionising radiations tend to use 2D detectors in order to improve the radiation detection efficiency. Then, fully 3D image reconstruction algorithms are required to recover the 3D image of the region of interest. These systems include 3D radiology, 3D X-ray computerised tomography, single photon emission computerised tomography (SPECT), and positron emission tomography (PET). The material is this text divided into four parts covering the following topics: cone-beam and new geometries reconstruction, SPECT quantitation, patient motion and gated SPECT, and PET quantitation and reconstruction.Aud ience:This work will be of interest to scientists, physicists and physicians seeking new information and insight in the on-going research work in this expanding field.

Appendix 164 3. A 3. A. 1 Approximate Estimation of Fundamental Matrix from General Matrix 164 3. A. 2 Estimation of Affine Transformation 165 4 RECOVERY OF EPIPOLAR GEOMETRY FROM LINE SEGMENTS OR LINES 167 Line Segments or Straight Lines 168 4. 1 4. 2 Solving Motion Using Line Segments Between Two Views 173 4. 2. 1 Overlap of Two Corresponding Line Segments 173 Estimating Motion by Maximizing Overlap 175 4. 2. 2 Implementation Details 4. 2. 3 176 Reconstructing 3D Line Segments 4. 2. 4 179 4. 2. 5 Experimental Results 180 4. 2. 6 Discussions 192 4. 3 Determining Epipolar Geometry of Three Views 194 4. 3. 1 Trifocal Constraints for Point Matches 194 4. 3. 2 Trifocal Constraints for Line Correspondences 199 4. 3. 3 Linear Estimation of K, L, and M Using Points and Lines 200 4. 3. 4 Determining Camera Projection Matrices 201 4. 3. 5 Image Transfer 203 4. 4 Summary 204 5 REDEFINING STEREO, MOTION AND OBJECT RECOGNITION VIA EPIPOLAR GEOMETRY 205 5. 1 Conventional Approaches to Stereo, Motion and Object Recognition 205 5. 1. 1 Stereo 205 5. 1. 2 Motion 206 5. 1. 3 Object Recognition 207 5. 2 Correspondence in Stereo, Motion and Object Recognition as 1D Search 209 5. 2. 1 Stereo Matching 209 xi Contents 5. 2. 2 Motion Correspondence and Segmentation 209 5. 2. 3 3D Object Recognition and Localization 210 Disparity and Spatial Disparity Space 210 5.

The concept of ridges has appeared numerous times in the image processing liter­ ature. Sometimes the term is used in an intuitive sense. Other times a concrete definition is provided. In almost all cases the concept is used for very specific ap­ plications. When analyzing images or data sets, it is very natural for a scientist to measure critical behavior by considering maxima or minima of the data. These critical points are relatively easy to compute. Numerical packages always provide support for root finding or optimization, whether it be through bisection, Newton's method, conjugate gradient method, or other standard methods. It has not been natural for scientists to consider critical behavior in a higher-order sense. The con­ cept of ridge as a manifold of critical points is a natural extension of the concept of local maximum as an isolated critical point. However, almost no attention has been given to formalizing the concept. There is a need for a formal development. There is a need for understanding the computation issues that arise in the imple­ mentations. The purpose of this book is to address both needs by providing a formal mathematical foundation and a computational framework for ridges. The intended audience for this book includes anyone interested in exploring the use­ fulness of ridges in data analysis.

This text covers Gaussian scale-space theory from its applications to its mathematical foundation. The reader not so familiar with scale-space should find it instructive to first consider some potential applications described in Part I. The next two parts both address fundamental aspects of scale-space. Whereas scale is treated as an essentially arbitrary constant in Part II, Part III emphasises the deep structure, i.e. the structure that is revealed by varying scale. Finally Part IV is devoted to non-linear extensions, notably non-linear diffusion techniques and morphological scale-spaces, and their relation to the linear case. The book is addressed to researchers in the field of image analysis seeking mathematical foundation of algorithms.

Motion-based recognition deals with the recognition of an object and/or its motion, based on motion in a series of images. In this approach, a sequence containing a large number of frames is used to extract motion information. The advantage is that a longer sequence leads to recognition of higher level motions, like walking or running, which consist of a complex and co-ordinated series of events. Unlike much previous research in motion, this approach does not require explicit reconstruction of shape from the images prior to recognition. This book provides a state-of-the-art study of this rapidly developing discipline. It consists of a collection of invited chapters by researchers throughout the world covering various aspects of motion-based recognition including lipreading, gesture recognition, facial expression recognition, gait analysis, cyclic motion detection, and activity recognition. This volume should be of interest to researchers and postgraduate students whose work involves computer vision, robotics and image processing.

Despite the fact that images constitute the main objects in computer vision and image analysis, there is remarkably little concern about their actual definition. In this book a complete account of image structure is proposed in terms of rigorously defined machine concepts, using basic tools from algebra, analysis, and differential geometry. Machine technicalities such as discretisation and quantisation details are de-emphasised, and robustness with respect to noise is manifest. From the foreword by Jan Koenderink: `It is my hope that the book will find a wide audience, including physicists - who still are largely unaware of the general importance and power of scale space theory, mathematicians - who will find in it a principled and formally tight exposition of a topic awaiting further development, and computer scientists - who will find here a unified and conceptually well founded framework for many apparently unrelated and largely historically motivated methods they already know and love. The book is suited for self-study and graduate courses, the carefully formulated exercises are designed to get to grips with the subject matter and prepare the reader for original research.'

One of the most natural representations for modelling spatial objects in computers is discrete representations in the form of a 2D square raster and a 3D cubic grid, since these are naturally obtained by segmenting sensor images. However, the main difficulty is that discrete representations are only approximations of the original objects, and can only be as accurate as the cell size allows. If digitization is done by real sensor devices, then there is the additional difficulty of sensor distortion. To overcome this, digital shape features must be used that abstract from the inaccuracies of digital representation. In order to ensure the correspondence of continuous and digital features, it is necessary to relate shape features of the underlying continuous objects and to determine the necessary resolution of the digital representation. This volume gives an overview and a classification of the actual approaches to describe the relation between continuous and discrete shape features that are based on digital geometric concepts of discrete structures.This book should be of interest to researchers and graduate students whose work involves computer vision, image processing, knowledge representation or representation of spatial objects.

This book contains the proceedings of the International Symposium on Mathematical Morphology and its Applications to Image and Signal Processing IV, held June 3-5, 1998, in Amsterdam, The Netherlands. The purpose of the work is to provide the image analysis community with a sampling of recent developments in theoretical and practical aspects of mathematical morphology and its applications to image and signal processing. Among the areas covered are: digitization and connectivity, skeletonization, multivariate morphology, morphological segmentation, color image processing, filter design, gray-scale morphology, fuzzy morphology, decomposition of morphological operators, random sets and statistical inference, differential morphology and scale-space, morphological algorithms and applications. Audience: This volume will be of interest to research mathematicians and computer scientists whose work involves mathematical morphology, image and signal processing.

In June 1998 the Fourth International Workshop on Digital Mammography was held in Nijmegen, The Netherlands, where it was hosted by the department of Radiology of the University Hospital Nijmegen. This series of meetings was initiated at the 1993 SPIE Biomedical Image Processing Conference in San Jose, USA, where a number of sessions were entirely devoted to mammographic image analysis. At very successful subsequent workshops held in York, UK (1994) and Chicago, USA (1996), the scope of the conference was broadened, establishing a platform for presentation and discussion of new developments in digital mammog­ raphy. Topics that are addressed at these meetings are computer-aided diagnosis, image processing, detector development, system design, observer performance and clinical evaluation. The goal is to bring researchers from universities, breast cancer experts, and engineers together, to exchange information and present new scientific developments in this rapidly evolving field. This book contains all the scientific papers and posters presented at the work­ shop in Nijmegen. Contributions came from as many as 20 different countries and 190 participants attended the meeting. At a technical exhibit companies demon­ strated new products and work in progress. Abstracts of all papers were reviewed by members of the scientific committee. Many of the accepted papers had excellent quality, but due to limited space not all of them could be included as full papers in these proceedings. Papers that were rated high by the reviewers are included as long or short papers, others appear as extended abstracts in the last chapter.

This text presents techniques for describing image textures. Contrary to the usual practice of embedding the images to known modelling frameworks borrowed from statistical physics or other domains, this book deduces the Gibbs models from basic image features and tailors the modelling framework to the images. This approach results in more general Gibbs models than can be either Markovian or non-Markovian and possess arbitrary interaction structures and strengths. The book presents computationally feasible algorithms for parameter estimation and image simulation and demonstrates their abilities and limitations by numerous experimental results. The book avoids too abstract mathematical constructions and gives explicit image-based explanations of all the notions involved.

This edited volume addresses a subject which has been discussed inten­ sively in the computer vision community for several years. Performance characterization and evaluation of computer vision algorithms are of key importance, particularly with respect to the configuration of reliable and ro­ bust computer vision systems as well as the dissemination of reconfigurable systems in novel application domains. Although a plethora of literature on this subject is available for certain' areas of computer vision, the re­ search community still faces a lack of a well-grounded, generally accepted, and--eventually-standardized methods. The range of fundamental problems encoIl!passes the value of synthetic images in experimental computer vision, the selection of a representative set of real images related to specific domains and tasks, the definition of ground truth given different tasks and applications, the design of experimental test­ beds, the analysis of algorithms with respect to general characteristics such as complexity, resource consumption, convergence, stability, or range of admissible input data, the definition and analysis of performance measures for classes of algorithms, the role of statistics-based performance measures, the generation of data sheets with performance measures of algorithms sup­ porting the system engineer in his configuration problem, and the validity of model assumptions for specific applications of computer vision.

A representation of objects by their parts is the dominant strategy for representing complex 3D objects in many disciplines. In computer vision and robotics, superquadrics are among the most widespread part models. Superquadrics are a family of parametric models that cover a wide variety of smoothly changing 3D symmetric shapes, which are controlled with a small number of parameters and which can be augmented with the addition of global and local deformations. The book covers, in depth, the geometric properties of superquadrics. The main contribution of the book is an original approach to the recovery and segmentation of superquadrics from range images. Several applications of superquadrics in computer vision and robotics are thoroughly discussed and, in particular, the use of superquadrics for range image registration is demonstrated. Audience: The book is intended for readers of all levels who are familiar with and interested in computer vision issues.

This is a comprehensive treatment of the extraction of landmarks from multi-modality images and the use of these features for elastic image registration. The emphasis is on model-based approaches, that is, the use of explicitly represented knowledge in computer vision. Both geometric models (describing the shape of objects) and intensity models (directly representing the image intensities) are utilized. The work describes theoretical foundations, computational and algorithmic issues, as well as practical applications, notably in medicine (neurosurgery and radiology), remote sensing, and industrial automation. Connections with computer graphics and artificial intelligence are illustrated.

Mathematical morphology is a powerful methodology for the processing and analysis of geometric structure in signals and images. This book contains the proceedings of the 5th International Symposium on Mathematical Morphology and its Applications to Image and Signal Processing, held June 26-28, 2000, at Xerox PARC, Palo Alto, California. It provides a broad sampling of theoretical and practical developments of mathematical morphology and its applications to image and signal processing. Areas covered include: decomposition of structuring functions and morphological operators, morphological discretization, filtering, connectivity and connected operators, morphological shape analysis and interpolation, texture analysis, morphological segmentation, morphological multiresolution techniques and scale-spaces, and morphological algorithms and applications. The subject matter of this volume should be useful to electrical engineers, computer scientists and mathematicians whose research work is focused on the theoretical and practical aspects of non-linear signal and image processing. It should also be of interest to those working in computer vision, applied mathematics and computer graphics.

Mathematical morphology (MM) is a powerful methodology for the quantitative analysis of geometrical structures. It consists of a broad and coherent collection of theoretical concepts, nonlinear signal operators, and algorithms aiming at extracting, from images or other geometrical objects, information related to their shape and size. Its mathematical origins stem from set theory, lattice algebra, and integral and stochastic geometry. MM was initiated in the late 1960s by G. Matheron and J. Serra at the Fontainebleu School of Mines in France. Originally it was applied to the the analysis of images from geological or biological specimens. However, its rich theoretical framework, algorithmic efficiency, easy implementability on special hardware, and suitability for many shape-oriented problems have propelled its widespread diffusion and adoption by academic and industry groups in many countries as one among the dominant image analysis methodologies.The purpose of this work is to provide the image analysis community with a sampling from the current developments in the theoretical (deterministic and stochastic) and computational aspects of MM and its applications to image and signal processing. The book consists of the papers presented at the ISMM'96 grouped into the following themes: + Theory + Connectivity + Filtering + Nonlinear System Related to Morphology + Algorithms/Architectures + Granulometries, Texture Segmentation + Image Sequence Analysis Learning + Document Analysis Applications.

The recent emergence of Local Binary Patterns (LBP) has led to significant progress in applying texture methods to various computer vision problems and applications. The focus of this research has broadened from 2D textures to 3D textures and spatiotemporal (dynamic) textures. Also, where texture was once utilized for applications such as remote sensing, industrial inspection and biomedical image analysis, the introduction of LBP-based approaches have provided outstanding results in problems relating to face and activity analysis, with future scope for face and facial expression recognition, biometrics, visual surveillance and video analysis. Computer Vision Using Local Binary Patterns provides a detailed description of the LBP methods and their variants both in spatial and spatiotemporal domains. This comprehensive reference also provides an excellent overview as to how texture methods can be utilized for solving different kinds of computer vision and image analysis problems. Source codes of the basic LBP algorithms, demonstrations, some databases and a comprehensive LBP bibliography can be found from an accompanying web site. Topics include: local binary patterns and their variants in spatial and spatiotemporal domains, texture classification and segmentation, description of interest regions, applications in image retrieval and 3D recognition - Recognition and segmentation of dynamic textures, background subtraction, recognition of actions, face analysis using still images and image sequences, visual speech recognition and LBP in various applications.Written by pioneers of LBP, this book is an essential resource for researchers, professional engineers and graduate students in computer vision, image analysis and pattern recognition. The book will also be of interest to all those who work with specific applications of machine vision.

Despite their novelty, wavelets have a tremendous impact on a number of modern scientific disciplines, particularly on signal and image analysis. Because of their powerful underlying mathematical theory, they offer exciting opportunities for the design of new multi-resolution processing algorithms and effective pattern recognition systems. This book provides a much-needed overview of current trends in the practical application of wavelet theory. It combines cutting edge research in the rapidly developing wavelet theory with ideas from practical signal and image analysis fields. Subjects dealt with include balanced discussions on wavelet theory and its specific application in diverse fields, ranging from data compression to seismic equipment. In addition, the book offers insights into recent advances in emerging topics such as double density DWT, multiscale Bayesian estimation, symmetry and locality in image representation, and image fusion. Audience: This volume will be of interest to graduate students and researchers whose work involves acoustics, speech, signal and image processing, approximations and expansions, Fourier analysis, and medical imaging.

Images and video play a crucial role in visual information systems and multimedia. There is an extraordinary number of applications of such systems in entertainment, business, art, engineering, and science. Such applications often involved large image and video collections, and therefore, searching for images and video in large collections is becoming an important operation. Because of the size of such databases, efficiency is crucial. We strongly believe that image and video retrieval need an integrated approach from fields such as image processing, shape processing, perception, database indexing, visualization, and querying, etc. This book contains a selection of results that was presented at the Dagstuhl Seminar on Content-Based Image and Video Retrieval, in December 1999. The purpose of this seminar was to bring together people from the various fields, in order to promote information exchange and interaction among researchers who are interested in various aspects of accessing the content of image and video data. The book provides an overview of the state of the art in content-based image and video retrieval. The topics covered by the chapters are integrated system aspects, as well as techniques from image processing, computer vision, multimedia, databases, graphics, signal processing, and information theory. The book will be of interest to researchers and professionals in the fields of multimedia, visual information (database) systems, computer vision, and information retrieval.

This volume describes the latest imaging and image analysis techniques that have been developed at leading centres for the visualization, analysis and understanding of normal and abnormal cardiac motion with magnetic resonance imaging (MRI). The use of MRI in measuring cardiac motion is particularly important because MRI is non-invasive, and it is the only modality capable of imaging detailed intramural motion within the myocardium. Biomedical engineers, medical physicists, computer scientists and physicians interested in learning about the latest advances in cardiovascular MRI should find this book to be a useful educational resource. Practitioners and researchers working in the field of cardiovascular MRI should find that the book provides practical technical details and references to other work, enabling the implementation of existing methods and serving as a basis for further research in the area.

This monograph explores the close relationship of various well-known pattern recognition problems that have so far been considered independent. These relationships became apparent with the discovery of formal procedures for addressing known problems and their generalisations. The generalised problem formulations were analysed mathematically and unified algorithms were found. The main scientific contribution of this book is the unification of two main streams in pattern recognition - statistical and structural. The material is presented in the form of ten lectures, each of which concludes with a discussion with a student.

MPEG-7 is the first international standard which contains a number of key techniques from Computer Vision and Image Processing. The Curvature Scale Space technique was selected as a contour shape descriptor for MPEG-7 after substantial and comprehensive testing, which demonstrated the superior performance of the CSS-based descriptor.

From the foreword by Thomas Huang: "During the past decade, researchers in computer vision have found that probabilistic machine learning methods are extremely powerful. This book describes some of these methods. In addition to the Maximum Likelihood framework, Bayesian Networks, and Hidden Markov models are also used. Three aspects are stressed: features, similarity metric, and models. Many interesting and important new results, based on research by the authors and their collaborators, are presented.Although this book contains many new results, it is written in a style that suits both experts and novices in computer vision."

Many approaches have been proposed to solve the problem of finding the optic flow field of an image sequence. Three major classes of optic flow computation techniques can discriminated (see for a good overview Beauchemin and Barron IBeauchemin19951): gradient based (or differential) methods; phase based (or frequency domain) methods; correlation based (or area) methods; feature point (or sparse data) tracking methods; In this chapter we compute the optic flow as a dense optic flow field with a multi scale differential method. The method, originally proposed by Florack and Nielsen [Florack1998a] is known as the Multiscale Optic Flow Constrain Equation (MOFCE). This is a scale space version of the well known computer vision implementation of the optic flow constraint equation, as originally proposed by Horn and Schunck [Horn1981]. This scale space variation, as usual, consists of the introduction of the aperture of the observation in the process. The application to stereo has been described by Maas et al. [Maas 1995a, Maas 1996a]. Of course, difficulties arise when structure emerges or disappears, such as with occlusion, cloud formation etc. Then knowledge is needed about the processes and objects involved. In this chapter we focus on the scale space approach to the local measurement of optic flow, as we may expect the visual front end to do. 17. 2 Motion detection with pairs of receptive fields As a biologically motivated start, we begin with discussing some neurophysiological findings in the visual system with respect to motion detection.

Many approaches have been proposed to solve the problem of finding the optic flow field of an image sequence. Three major classes of optic flow computation techniques can discriminated (see for a good overview Beauchemin and Barron IBeauchemin19951): gradient based (or differential) methods; phase based (or frequency domain) methods; correlation based (or area) methods; feature point (or sparse data) tracking methods; In this chapter we compute the optic flow as a dense optic flow field with a multi scale differential method. The method, originally proposed by Florack and Nielsen [Florack1998a] is known as the Multiscale Optic Flow Constrain Equation (MOFCE). This is a scale space version of the well known computer vision implementation of the optic flow constraint equation, as originally proposed by Horn and Schunck [Horn1981]. This scale space variation, as usual, consists of the introduction of the aperture of the observation in the process. The application to stereo has been described by Maas et al. [Maas 1995a, Maas 1996a]. Of course, difficulties arise when structure emerges or disappears, such as with occlusion, cloud formation etc. Then knowledge is needed about the processes and objects involved. In this chapter we focus on the scale space approach to the local measurement of optic flow, as we may expect the visual front end to do. 17. 2 Motion detection with pairs of receptive fields As a biologically motivated start, we begin with discussing some neurophysiological findings in the visual system with respect to motion detection.

Mathematical Morphology is a speciality in Image Processing and Analysis, which considers images as geometrical objects, to be analyzed through their interactions with other geometrical objects. It relies on several branches of mathematics, such as discrete geometry, topology, lattice theory, partial differential equations, integral geometry and geometrical probability. It has produced fast and efficient algorithms for computer analysis of images, and has found applications in bio-medical imaging, materials science, geoscience, remote sensing, quality control, document processing and data analysis.This book contains the 43 papers presented at the 7th International Symposium on Mathematical Morphology, held in Paris on April 18-20, 2005. It gives a lively state of the art of current research topics in this field. It also marks a milestone, the 40 years of uninterrupted development of this ever-expanding domain.

Computer vision and image analysis require interdisciplinary collaboration between mathematics and engineering. This book addresses the area of high-accuracy measurements of length, curvature, motion parameters and other geometrical quantities from acquired image data. It is a common problem that these measurements are incomplete or noisy, such that considerable efforts are necessary to regularise the data, to fill in missing information, and to judge the accuracy and reliability of these results. This monograph brings together contributions from researchers in computer vision, engineering and mathematics who are working in this area. The book can be read both by specialists and graduate students in computer science, electrical engineering or mathematics who take an interest in data evaluations by approximation or interpolation, in particular data obtained in an image analysis context.

This book traces progress in photography since the first pinhole, or camera obscura, architecture. The authors describe innovations such as photogrammetry, and omnidirectional vision for robotic navigation. The text shows how new camera architectures create a need to master related projective geometries for calibration, binocular stereo, static or dynamic scene understanding. Written by leading researchers in the field, this book also explores applications of alternative camera architectures.