Computer Perceptual Organization In Computer Vision

This book describes the design of a complete, flexible system for perceptual organization in computer vision using graph theoretic techniques, voting methods, and an extension of the Bayesian networks called perceptual inference networks (PINs). The PIN, which forms the heart of the system and which is based on Bayesian probabilistic networks, exhibits potential for application in several areas of computer vision as well as a range of other spatial reasoning tasks. The text includes a highly comprehensive, classificatory review of prior work in perceptual organization and, within that framework, identifies key areas for future work by the computer vision research community.

Perceptual Organization for Artificial Vision Systems is an edited collection of invited contributions based on papers presented at The Workshop on Perceptual Organization in Computer Vision, held in Corfu, Greece, in September 1999. The theme of the workshop was `Assessing the State of the Community and Charting New Research Directions.' Perceptual organization can be defined as the ability to impose structural regularity on sensory data, so as to group sensory primitives arising from a common underlying cause. This book explores new models, theories, and algorithms for perceptual organization. Perceptual Organization for Artificial Vision Systems includes contributions by the world's leading researchers in the field. It explores new models, theories, and algorithms for perceptual organization, as well as demonstrates the means for bringing research results and theoretical principles to fruition in the construction of computer vision systems. The focus of this collection is on the design of artificial vision systems. The chapters comprise contributions from researchers in both computer vision and human vision.

COMPUTER VISION is a field of research that encompasses many objectives. A primary goal has been to construct visual sensors that can provide general-purpose robots with the same information about their surroundings as we receive from our own visual senses. This book takes an important step towards this goal by describing a working computer vision system named SCERPO. This system can recognize known three-dimensional objects in ordinary black-and-white images taken from unknown viewpoints, even when parts of the object are undetectable or hidden from view. A second major goal of computer vision re search is to provide a computational understanding of human vision. The research presented in this book has many implica tions for our understanding of human vision, particularly in the areas of perceptual organization and knowledge-based recogni tion. An attempt has been made to relate each computational result to the relevant areas in the psychology of vision. Since the material is meant to be accessible to a wide range of inter disciplinary readers, the book is written in plain language and attempts to explain most concepts from the starting position of the non-specialist. vii viii PREFACE One of the most important conclusions ansmg from this research is that visual recognition can commonly be achieved directly from the two-dimensional image without any prelim inary reconstruction of depth information or surface orienta tion from the visual input.

Introduction -- Tensor voting -- Stereo vision from a perceptual organization perspective -- Tensor voting in ND -- Dimensionality estimation manifold learning and function approximation -- Boundary inference -- Figure completion -- Conclusions -- References.

This lecture presents research on a general framework for perceptual organization that was conducted mainly at the Institute for Robotics and Intelligent Systems of the University of Southern California. It is not written as a historical recount of the work, since the sequence of the presentation is not in chronological order. It aims at presenting an approach to a wide range of problems in computer vision and machine learning that is data-driven, local and requires a minimal number of assumptions. The tensor voting framework combines these properties and provides a unified perceptual organization methodology applicable in situations that may seem heterogeneous initially. We show how several problems can be posed as the organization of the inputs into salient perceptual structures, which are inferred via tensor voting. The work presented here extends the original tensor voting framework with the addition of boundary inference capabilities; a novel re-formulation of the framework applicable to high-dimensional spaces and the development of algorithms for computer vision and machine learning problems. We show complete analysis for some problems, while we briefly outline our approach for other applications and provide pointers to relevant sources.

This book considers computer vision to be an integral part of the artificial intelligence system. The core of the book is an analysis of possible approaches to the creation of artificial vision systems, which simulate human visual perception. Much attention is paid to the latest achievements in visual psychology and physiology, the description of the functional and structural organization of the human perception mechanism, the peculiarities of artistic perception and the expression of reality. Computer vision models based on these data are investigated. They include the processes of external data analysis, internal environmental model synthesis, and the generating of behavioristic responses based on external and internal models comparison. Computer vision system evolution resulting from environmental effects is also considered. A unique feature of this book is the authors' use of black and white, and colour prints of traditional and contemporary Russian art to illustrate their principal theses. In doing so, they introduce the reader to a particularly Russian view of the world.

Structured Computer Vision