From Fragments to Objects

Segmentation and Grouping in Vision

Edited by

  • Thomas F. Shipley, Temple University
  • Philip J. Kellman, University of California, Los Angeles

The simple act of seeing an object conceals an array of fascinating and difficult scientific questions. Because objects occlude themselves and each other, the information available for seeing objects is fragmentary. Yet we perceive whole objects, and doing so provides a key foundation for action, thought, language and learning. As we move, the visible parts of objects change, yet we experience these objects as being stable over time. The phenomenal coherence and stability of objects are simple facts of perceptual experience, but they result from complex computational processes. Despite years of work, researchers in biological and computational vision have yet to achieve satisfactory accounts of unit formation. The present period, however, is one of new insights and rapid progress.
This book addresses the problem of how the human visual system organizes inputs that are fragmented in space and time into coherent, stable perceptual units - objects. In doing so it addresses the following questions: what kinds of segmentation and grouping abilities exist in human perceivers? What information and computational processes achieve segmentation and grouping? What are the psychological consequences of perceiving whole objects?
In an effort to give a comprehensive, integrative answer to these questions the volume includes chapters from authors in five areas: philosophical foundations; computational and neural models of segmentation and grouping; attention and grouping; development; segmentation and grouping over time (event perception).
The chapters review the current state of the field, describe recent theoretical and empirical research, and provide some explicit suggestions for future research direction.
From Fragments to Objects: Segmentation and Grouping in Vision takes a comprehensive cognitive science approach to object perception, brings together separate lines of research in object perception in one volume, gives an integrated and up-to-date review of theory and empirical research and offers directions for future study.
View full description


Book information

  • Published: November 2001
  • Imprint: NORTH-HOLLAND
  • ISBN: 978-0-444-50506-4

Table of Contents

Acknowledgements. List of contributors. Philosophy and History of Perceptual Unit Formation. The concept of an "object" in perception and cognition (R. Schwartz). Balls of wax and cans of worms: the early history of object perception (M. Atherton). Development. Perceptual unit formation in infancy. (M.E. Arterberry). Perceptual units and their mapping with language (B. Landau). Attention. An object substitution theory of visual masking (J.T. Enns, V. Di Lollo). Attention and unit formation: a biased competition account of object-based attention (S.P. Vecera, M. Behrmann). Models of Segmentation and Grouping. Geometric and neural models of object perception (P.J. Kellman et al.). Varieties of grouping and its role in determining surface layout (B. Gillam). Amodal completion: a case study in grouping (A.B. Sekuler, R.F. Murray). Perceptual organization as generic object recognition (D.W. Jacobs). Simplicity, regularity, and perceptual interpretations: a structural information approach (R. Van Lier). Computational neural models of spatial integration in perceptual grouping (H. Neumann, E. Mingolla). Part-based representations of visual shape and implications for visual cognition (M. Singh, D.D. Hoffman). Spatiotemporal Segmentation and Grouping. Gaze control for face learning and recognition by humans and machines (J. Henderson et al.). The visual interpretation of object and human movement (M. Shiffrar). Contours from apparent motion: a computational theory (W. Prophet et al.). Breathing illusions and boundary formation in space-time (N. Bruno). Perception of occluding and occluded objects over time: spatiotemporal segmentation and unit formation (T.F. Shipley, D.W. Cunningham). Author index. Subject index.